Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: emerson909 on October 01, 2024, 11:44:42 am
-
Thought I would post some details of my second build, I got a hot rod deluxe for a good price and after the success of my Bassman conversion decided to have a go at a single channel AB763. It was quite tight for space to add additional preamp valves so only added 1 for the reverb and went for a MOSFET driven bias wiggle tremolo. I also wanted to use the high impedance tank that came with the HRD so went for Merlins transformer less reverb from the valve wizard website. These were both late changes after I had built the turret board as well as switching to radial filter caps so the final layout is a bit congested in the preamp but works ok. I’ve kept the HRD 4K:8 ohm output transformer so using the 4ohm tap with an 8 ohm speaker and JJ 6V6’s. While reading about this I noticed I had installed the NFB resistor values for the multi speaker AB763s so switched the feedback to the 4 ohm tapping and kept the resistors.
The build went well but the start up has been a somewhat humbling experience :BangHead: I’ll list some of the issues I had and the fixes in case it might help anybody or if anybody has any comments as to what I could have done better! I don’t have a scope so the troubleshooting has been a bit hit and miss.
I had an oscillation on initial power up leading to a high pitch squeal, no problem I thought, swapped the OT primaries and the squeal went away until I maxed out all the controls. Once the treble was on full I had another oscillation leading to loud squeal. Lots of swapping valves, moving wires, checking solder joints etc but nothing seemed to make a difference so I put a 47pF cap across the LTP plates. That seemed to tame the oscillation but not eliminate it. Then came across a post by pdf64 suggesting caps across the preamp load resistors so put 47pF caps across V1A and B and the oscillation was gone!
I was then noticing a trailing distortion on the lower notes, lots of reading led me to think this was blocking distortion caused by another oscillation. Cue lots more valve swapping, bias changing, chop sticking, moving leads, resoldering etc but no improvements. I hooked up the speaker from my other amp and turned out it was the speaker. I should have picked this up far sooner as the output volume was also really low.
Swapped speakers and everything was sounding good until I turned the reverb up over 3. I would get a nasty ringing sound. I tried wrapping the tank in all sorts, moving it into a different room, put the bright cap on a switch, clipped the 10pF bypass across the 3.3M resister (no improvement to reverb but sounds better so kept). Put a 2200pF cap over the recovery grid leak, no improvement. Increased said cap to 0.1uF which I initially thought had worked but it had just rolled of the highs, so I now had a mid-range ringing.
I put 10K grid stoppers on all of the preamp valves which made no difference so on the suggestion of another forum post I removed the 100uF reverb driver cathode bypass which has stopped the ringing. I went back to the 2200pF grid leak bypass cap to brighten up the reverb, it’s not as deep as before but good enough for what I want. Not sure how I could tweak it to get slightly more depth?
The tremolo was also making a thumping sound so added a diode across the intensity pot as suggested by Slucky which got rid of the thumping.
Need to get a new speaker once funds allow but otherwise very pleased with the outcome.
-
Schematic
-
Pictures
-
...everything was sounding good until I turned the reverb up over 3. I would get a nasty ringing sound. I tried wrapping the tank in all sorts, moving it into a different room, put the bright cap on a switch, clipped the 10pF bypass across the 3.3M resister (no improvement to reverb but sounds better so kept). Put a 2200pF cap over the recovery grid leak, no improvement. Increased said cap to 0.1uF which I initially thought had worked but it had just rolled of the highs, so I now had a mid-range ringing.
I put 10K grid stoppers on all of the preamp valves which made no difference so on the suggestion of another forum post I removed the 100uF reverb driver cathode bypass which has stopped the ringing. I went back to the 2200pF grid leak bypass cap to brighten up the reverb, it’s not as deep as before but good enough for what I want. Not sure how I could tweak it to get slightly more depth?
You can try a 560pF plate to cathode cap with the cathode bypass cap restored to your reverb driver.
If that doesn't cure it you can try a series resistor with the cathode cap. Something in the 1k-3.3k range may be enough to increase depth without inducing oscillation.
Finally you can try a split plate load, say 6.8k and 3.3k for a cut of about a third of the voltage from the driver. Adjust to taste. 8.2k/1.8k would be another good one to try.
-
Thanks for the suggestions, I'll try crocodile clipping in some components and report back with the results.
-
... Once the treble was on full I had another oscillation leading to loud squeal. Lots of swapping valves, moving wires, checking solder joints etc but nothing seemed to make a difference so I put a 47pF cap across the LTP plates. That seemed to tame the oscillation but not eliminate it. Then came across a post by pdf64 suggesting caps across the preamp load resistors so put 47pF caps across V1A and B and the oscillation was gone!
My guess is that the convoluted layout may be causing an excessive degree of unintended coupling between early and later stages, forming feedback loops sufficient to result in free running oscillation. eg the tone stack components are mounted over by V3 and 4 components / circuits.
It's best for the layout to follow the schematic; the input and output sockets, valve sockets and front panel controls are part of that, it's not just the circuit board. But with any real world build, there's a compromise, hence it's best to follow known good chassis / and board layouts.
... everything was sounding good until I turned the reverb up over 3. I would get a nasty ringing sound. I tried wrapping the tank in all sorts, moving it into a different room, put the bright cap on a switch, clipped the 10pF bypass across the 3.3M resister (no improvement to reverb but sounds better so kept). Put a 2200pF cap over the recovery grid leak, no improvement. Increased said cap to 0.1uF which I initially thought had worked but it had just rolled of the highs, so I now had a mid-range ringing.
I put 10K grid stoppers on all of the preamp valves which made no difference so on the suggestion of another forum post I removed the 100uF reverb driver cathode bypass which has stopped the ringing. ...
The layout may be contributing to that too.
But grounding could be playing a part. I think R31, R32 and its bypass cap, and the return from the tank, need grounding to C36 B node cap -ve lead. ie don't let them ground out anywhere else.
Rather than completely remove the R32 bypass, try a resistor in series with it, start with 470R same as R32. Reduce it if that's totally stable, eg try 220R.
-
Thanks for the feedback (no pun!), massively appreciate it. Yes this might have to be my prototype board, switching to radial filter caps and moving the power tube screen resistors to the sockets freed up lots of space towards that end of the board. I also realised there's no need for the trem circuit to be where it is and it would be better off over by the bias circuit. I could then also swap the position of the reverb driver and recovery circuits so the wires aren't crossing. If I’d realised this all at once I would have redone the board but it sort of developed as I was going after I’d staked all of the turrets! I’ve really enjoyed building it so will be fun to have another go.
I moved the tonestack to where it is to keep the wires to the pots as short as possible after seeing the Psionic audio deluxe reverb PCB conversions.
it's best to follow known good chassis / and board layouts.
Yeah I started off based on the Hoffman single channel but then changed the trem and couldn’t find a layout for that circuit so had to go freestyle, then squeezed the reverb load resistor and that giant 1uF cap in :sad2:.
But grounding could be playing a part. I think R31, R32 and its bypass cap, and the return from the tank, need grounding to C36 B node cap -ve lead. ie don't let them ground out anywhere else.
I did also think this and separated out all the tremolo grounds but I had the reverb grounded to the preamp bus per the Hoffman design so it was going to be more difficult to separate and it went in the ‘if all else fails’ category! Maybe one to look at for Rev 2.
Rather than completely remove the R32 bypass, try a resistor in series with it, start with 470R same as R32. Reduce it if that's totally stable, eg try 220R.
This is the way I’m going to go, will clip some resistors in over the weekend and see how it goes. Thanks again!
-
Decided I'm going to redo the board as attached, hopefully this will get rid of some of the gremlins!
-
...
I did also think this and separated out all the tremolo grounds but I had the reverb grounded to the preamp bus per the Hoffman design so it was going to be more difficult to separate and it went in the ‘if all else fails’ category! Maybe one to look at for Rev 2.
...
The reverb drive circuit is very different though to the trad arrangement used with Hoffman, the transformer is a bandpass filter and provides galvanic isolation.
So for V2 keep it simple if you wish, but I suggest to make provision in the layout to keep things in line with Merlin's guidelines, eg add a few suitably placed turrets http://www.valvewizard.co.uk/Grounding.pdf
Also reading through that might explain why referring the bias winding to the HT rectifier negative lug doesn't seem good practice.
The V2 cathode bypass cap will need an extra turret for a series resistor?
-
So for V2 keep it simple if you wish
Ha ha no I thought I had separated them but looking at it yes can see there's a few issues, back to the drawing board!
Also reading through that might explain why referring the bias winding to the HT rectifier negative lug doesn't seem good practice.
I can't see what you mean on this one, I've got one side of the bias winding connected to D15 and 1 to the power section earth point (along with the HT rectifier negative). Not sure where else I would ground it, it seems to follow Merlin's scheme but maybe I'm missing something??
The V2 cathode bypass cap will need an extra turret for a series resistor?
Yes I've got this one although just got a link in for now as hopefully I won't need it
Thanks again!
-
With regards the reverb I've moved R31 and R32 to B node ground. Presumably R94 and R33 are part of the input to V3A and should be grounded to the D node per the attached schematic?
I hadn't really realised until you mentioned it earlier regarding the transformer providing isolation between the stages, think I can see the issues a bit better now.
-
So you're still using the original reverb tank?
If so, you're trying to build a tube reverb into a tank that was designed for transistors.
There was a thread recently regarding a transformerless reverb, and there are different driver tubes that work better with different impedences, and there may be options for building a voltage divider that works with your tank, but that would take far more intelligent people than me.
The end result in that thread turned out that the best solution was a new tank. Coupled with a certain driver and voltage divider.
-
So you're still using the original reverb tank?
Hi Al, yes it is the original 600 ohm tank, I've followed Merlin Blencow's design here;https://www.valvewizard.co.uk/reverbdriver.html (https://www.valvewizard.co.uk/reverbdriver.html)
I've measured 4.9V across the 12AT7 470R cathode resistor so I think the 12AT7 is correctly biased for 10mA as needed.
I've been out with the soldering iron this morning and separated R31, R32 and it's bypass cap and R94 from the preamp ground bus. They're now grounded direct to chassis on one of the reverb jack tabs.
R33 was previously grounded to chassis but I've wired that to the preamp ground bus, I had to increase the bypass capacitor slightly to 3.3 nF as I couldn't reuse the 2.2 and didn't have another.
I've got a 470R wired in series with R32 cathode bias. Looks like the updated schematic attached.
Initial signs are promising, it's too early to test with any volume but with the channel volume up and master volume down I'm not getting the ringing where I would previously.
Will hopefully get a chance to do some mor testing later.
-
Based on info in a thread about cathode bypass caps and hum I'd recommend swapping the locations of the 100μF and the 22μF.
Also, there's good reason to believe the grid leak bypass cap forms an RLC resonance circuit with the reverb tank output transducer, and can actually peak the treble instead of cut it. If you want to shape the treble try leaving that cap out and instead place a 330-680pF cap across the plate resistor.
-
Thanks, the 560pf caps you recommended arrived earlier, managed to install one across the driver load resistor and I think the cumulative effect of today’s mods has worked. Still not had chance for a proper test but it’s definitely better than before. Thanks again
-
OK so got chance to test the amp this morning and I'm really happy with it, the reverb is totally usable and much deeper than with the cathode resistor unbypassed, I don't get the ringing at any realistic setting now.
I recorded a quick demo of the reverb and trem, I'm a terrible player so please try to ignore that!
https://www.dropbox.com/scl/fi/icv8zpd2ryt3gfm99bglu/AB763-reverb-and-trem.wav?rlkey=56tncb8ft46s2bjdaaeok7bh3&st=gpc476nc&dl=0 (https://www.dropbox.com/scl/fi/icv8zpd2ryt3gfm99bglu/AB763-reverb-and-trem.wav?rlkey=56tncb8ft46s2bjdaaeok7bh3&st=gpc476nc&dl=0)
I'm playing a les paul, bridge pickup into input 2 with volume around 2.5 and master volume on 10, reverb around 3.5
I'm still going to do a second version of the board but not in so much of a hurry now.
Thanks for everyone's help.
-
This is how the Reverb is wired now with the revised grounding and plate bypass cap installed
-
...
Also reading through that might explain why referring the bias winding to the HT rectifier negative lug doesn't seem good practice.
I can't see what you mean on this one, I've got one side of the bias winding connected to D15 and 1 to the power section earth point (along with the HT rectifier negative). Not sure where else I would ground it, it seems to follow Merlin's scheme but maybe I'm missing something??
From the Merlin grounding document -
Fig. 15.4: Current flows around the transformer-rectifier-reservoir in short heavy pulses that may be a source of buzz.
...
To minimise the interaction of ripple current with the rest of the circuit, the transformer-rectifier-reservoir must be treated as a single, self-contained circuit block.
In some schematics this is emphasised by drawing the circuit in a more compact fashion, as in fig. 15.5, and the circuit should be physically built in a similarly compact way.The rest of the amplifier will then be connected directly to the
terminals of the reservoir capacitor. Connections to any other points on this noisy current loop are not allowed, and no part of this network may be connected to the chassis if a quiet ground scheme is to be maintained!
So the reservoir cap negative terminal should be the local star point / start of the common bus.
Whereas your layout shows the rectifier negative terminal used for this.
Furthermore -
In a fixed-bias amp the negative bias supply should be considered to be a tiny power supply in its own right, so the same grounding logic can be applied. The whole bias supply is built with its own star (or possibly bus) ground scheme, and the last stage
-often a bias adjustment pot-is finally connected to the audio circuit, as in fig. 15.13.
-
With regards the reverb I've moved R31 and R32 to B node ground. Presumably R94 and R33 are part of the input to V3A and should be grounded to the D node per the attached schematic?
...
Sorry, it's probably not significant but R94 should be referred back to the B node cap too.
More significantly, so should the shield lug of the reverb drive RCA socket. ie to do that, insulated washers will be needed to prevent it connecting to the chassis.
Whereas R33 should be referred to the local D node star point, same as the cathode resistor of the stage it's related to.
Keep in mind the heavy-ish current loop driving the tank, and the miniscule current loop at the inputs.
I recorded a quick demo of the reverb and trem
Well done, it sounds great, I especially like the trem!
-
Right, I think I see what you mean, moving the existing wire from ground to rectifier to the filter cap ground (purple arrows in attachment). I was thinking both were electrically similar but not thinking about current flow. I can make that change pretty easily.
-
Right, I think I see what you mean, moving the existing wire from ground to rectifier to the filter cap ground (purple arrows in attachment). I was thinking both were electrically similar but not thinking about current flow. I can make that change pretty easily.
Yes, that's it :thumbsup:
-
Sorry, it's probably not significant but R94 should be referred back to the B node cap too.
More significantly, so should the shield lug of the reverb drive RCA socket. ie to do that, insulated washers will be needed to prevent it connecting to the chassis.
I've got both of these grounded to the chassis atm, it's going to be difficult to get the insulated washers on the RCA jack so will leave that for when I redo the board, there is a noticeable buzz on the reverb once it gets over midway so this could be the cause.
Well done, it sounds great, I especially like the trem!
Thanks!
-
Right, I think I see what you mean, moving the existing wire from ground to rectifier to the filter cap ground (purple arrows in attachment). I was thinking both were electrically similar but not thinking about current flow. I can make that change pretty easily.
Yes, that's it :thumbsup:
No, that's not it, it's close but not right.
That next filter cap C36 can't be directly attached to C34. Has to be a wire between the 2.
The way it's wired now, it's wired directly into that 1st filter cap loop.
-
ok will have to design that into the rev 2 board. I already re wired the existing amp! Looks like this now. It's really quiet with no signal, just a bit of hiss.
-
This is what Rev 2 is currently looking like incorporating previous comments,
I've hopefully now got reverb driver and trem circuits grounded to the B node which is separated with a bit of wire from the A node. Will add insulating washers to J8 when the board is out.
I also swapped the master volume and trem pots around to shorten the leads and avoid crossing.
-
The grounding seems a bit convoluted, doesn't make sense to me, eg the controls seem to be connected to the C node?
Have a look at how Michael at Modulus arranges things with local star points that get linked together and then the input stage cathode return is connected to the chassis near the input.
https://download1655.mediafire.com/uod54furnnjgbv1swTr78UYkbEvrefPdf9QtUJG37W6M2D_BzaKqw9iI0Yte4-RGoF1CZisldeLxEMQu263g3WYrZAFQaCZq4_xk_9BCYQ6YbMJLkHiaGeeEHEI2iO3yzn3kAReS2elFKEAO9YvKriMkpwa2T1F3JpIgTFU/hma7541vr8kou1z/GTS-20+Layout.pdf
I suppose as the current arrangement isn't hummy, it's a somewhat academic point :)
-
I'm would think his grounding scheme works, or he wouldn't post it, but,
He's got the heater CT connected with the 1st filter cap ground and PT B+ CT.
So the charging pulses can be pushed into the heaters wiring, I wouldn't do that.
-
Good point :thumbsup:
To eliminate the potential for heater buzz interference, I recommend DC elevation of the heater circuit.
But yeah, as it works fine as is, then theoretical improvements beyond that becomes an 'angels on a pin head' kinda thing.
There's nothing wrong with that though, provided we don't let 'best' become the enemy of 'good enough' :)
-
The grounding seems a bit convoluted, doesn't make sense to me, eg the controls seem to be connected to the C node?
I was thinking the same :icon_biggrin:
It might not be clear from the layout as to where the grounds actually are as I don't show the earths on the tag strips. Also not clear that the input jacks are isolated from the chassis. I've uploaded another layout showing the grounds and slightly rearranged preamp bus.
There are 4 connections to chassis;
1. Safety earth for 240VAC supply connected direct to chassis via bolt and star washer with nothing else.
2. Power Section earth for A and B node bus and bias 50VAC neutral connected to chassis via 1 of the tag strip grounded terminals.
3. Heater artificial centre tap connected to chassis via 1 of the tag strip grounded terminals.
4. C and D node bus connected to chassis via 1 of the tag strip grounded terminals near the input jacks. The other end of this bus is insulated from chassis.
I'm trying to follow the principle of having 2 main ground busses, It would be difficult to implement a scheme like Michael's as I'd have around 9 terminations to make on the pre amp filter cap -ve terminal, so would end up having to do some sort of bus arrangement anyway.
I noticed I had sort of run another bus for the preamp in the last layout I posted, I've split this up for each stage now with a single connection to the main bus in circuit order per Merlin's 'improved bus' scheme
As you have pointed out I think I'm getting to the point of diminishing returns as it's already a quiet amp although I obviously want to make it as good as I practically can. I think elevating the heater currents will be one for a future build :)
Thanks once again for all the feedback, you've all really made me think about some aspects of the design!
-
There's nothing wrong with that though, provided we don't let 'best' become the enemy of 'good enough' :)
Just have to move that 1 wire. :dontknow:
-
Schematic
I want this schematic close, so we don't have to scroll all the way back to almost the start. :laugh:
https://el34world.com/Forum/index.php?action=dlattach;topic=32112.0;attach=116012
-
Had to make a small change, I had the tremolo supply connected to B+.
-
...
I'm trying to follow the principle of having 2 main ground busses, It would be difficult to implement a scheme like Michael's as I'd have around 9 terminations to make on the pre amp filter cap -ve terminal, so would end up having to do some sort of bus arrangement anyway.
I noticed I had sort of run another bus for the preamp in the last layout I posted, I've split this up for each stage now with a single connection to the main bus in circuit order per Merlin's 'improved bus' scheme
To continue discussing how to make an amp that doesn't hum less hummy :laugh: -
Rather than Merlin's improved bus, by having the separate power amp and preamp connections to the chassis arrangement, what's happening is that charging / ripple current for the C and D nodes is passing from the reservoir, via the chassis, through the preamp chassis connection, to the caps. See the red ripply current path below :)
(https://i.ibb.co/KGHpFWF/IMG-3776.jpg)
I suggest that to better approximate Merlin's layout, you lift the power amp, reservoirs etc off the chassis, and make a wire link between the B and C cap negative terminals.
-
To continue discussing how to make an amp that doesn't hum less hummy :laugh: -
Yes we will get rid of this hum whether it exists or not! :icon_biggrin:
-
Finished the Rev 2 build, the new layout definitely makes more sense, I'm away for a week from tomorrow so not going to get chance to test it until I get back, will let you how it goes.
-
Well this has turned out to be a complete disaster, I think I‘ve built a Geiger counter with added squeal at max vol and treble. :BangHead:
Amp is just a complete mess of squeal, static, hiss and ironically hum! Video of the horror show;
https://www.dropbox.com/scl/fi/s1qxbo4tnk2ovx133ww8u/20241103_192915000_iOS.MOV?rlkey=3t0jhdrdcovjze78j8h5pn1v4&st=8zhxs02m&dl=0 (https://www.dropbox.com/scl/fi/s1qxbo4tnk2ovx133ww8u/20241103_192915000_iOS.MOV?rlkey=3t0jhdrdcovjze78j8h5pn1v4&st=8zhxs02m&dl=0)
The static is not affected by the channel volume, but when the reverb is engaged and turned up the static goes??
To try to isolate the issue I’ve done the following;
Chopsticking around V3B grid seemed to have an impact so removed the grid stopper and remade the connections at the valve and board but no change.
Turned off trem and reverb using footswitch
Remove reverb driver (V2) – This gets rid of the oscillation but the static and hiss remain.
Removed V4 (PI), reduces volume of static but still present.
Removed V3, no change
Removed V1, no change
Reinstated V1, V3,
Grounded V1A grid, no change
Grounded V1B grid, no change
Grounded V3A grid, no change
Grounded V3B grid, big reduction in static
As volume has no impact assume V1A is good.
Lifted V1B plate resistor, no change
Lifted V3A plate resistor, no change
Lifted V3B plate resistor, big reduction in static
Replaced V3B load resistor, no change
Replaced R37, R38, no change
Replaced V2 plate resistor R92 and R93, R94, no change
Changed ground scheme to pdf64 last suggestion so A and B node is off chassis and connected to end of pre amp bus bar, effectively connected to C no change.
Seems the issue is something around V3B but I’m out of ideas now other than to go back to previous layout.
All valves and reverb tank were good before the amp was dismantled (destroyed).
Any help appreciated!! :help: :help:
-
You have 2 black wires attached to the 1st B+ filter cap ground lead. They go over to the corner of the chassis. 1 goes over close to the rectifier socket, I can't see what it's hooked up to?
The other goes over to the solder lug strip, and seems to connect with 2 off colored black/dark charcoal gray that come from the PT. What are those wires? Are they the -bias wind fly leads?
And 1 of them is grounded with the 1st B+ filter cap ground lead?
And that solder strip lug is bolted to the chassis for ground?
-
One goes to the rectifier -ve, one used to go to chassis ground although I moved it to the end of the preamp ground bus, moving it made no difference
-
And the 2nd B+ filter caps ground lead has a black wire attached to it coming from what looks like a chassis grommet hole that has OT wires coming through?
Where does that black wire go? What's it connected to?
-
One goes to the rectifier -ve, one used to go to chassis ground although I moved it to the end of the preamp ground bus, moving it made no difference.
This makes no sense.
What are those 2 off colored black/dark charcoal gray that come from the PT?
-
2 off colored black/dark charcoal gray that come from the PT. What are
And that solder strip lug is bolted to the chassis for ground?
Off colour wires are the 220/230v taps off the pt, they are just terminated on spare terminals. I’m using the black 240v tapping.
Yes the tag strip has 2 grounded terminals, one has heater artificial centre tap and the other has the bias 50vac neutral, used to also have the b+ ground before I moved it
-
One goes to the rectifier -ve,
I don't understand this.
What rectifier negative?
And where's the PT's B+ high acv wind's CT?
-
Not sure how to explain it then but you can see it better in the pic in reply 23.
It’s the same as it was before when it was working
-
I'm asking about the wires at the end of the solder lug strip close to the fuse.
There's a brown wire there that goes from the solder lug strip over to the -bias circuit?
-
I'm asking about the 4 wires at the very end of that solder lug strip next to the fuse and next to/just below the red wires.
-
Ok, so this PT has no B+ CT?
Your using a SS FWB rectifier?
That's why you said rectifier negative.
-
I'm asking about the 4 wires at the very end of that solder lug strip next to the fuse and next to/just below the red wires.
The pt has no ht centre tap
It has a 50vac tap which is the brown wires. Hot is terminated into the 1st terminal and a wire is taken across to the bias rectifier diode
Neutral is terminated into the second grounded terminal. This used to also have the B+ (black wire) but I moved that to the preamp ground bus so now the second terminal only has the 50vac neutral.
-
Ok, so this PT has no B+ CT?
Your using a SS FWB rectifier?
That's why you said rectifier negative.
Yes correct, sorry we’re a bit out of sync with our replies! Trying to be clear
-
Its shown on page 3 of the schematic, reply 1
-
The pt has no ht centre tap
It has a 50vac tap which is the brown wires. Hot is terminated into the 1st terminal and a wire is taken across to the bias rectifier diode
Neutral is terminated into the second grounded terminal. This used to also have the B+ (black wire) but I moved that to the preamp ground bus so now the second terminal only has the 50vac neutral.
Ok.
This has to be changed.
-
Ok.
This has to be changed.
I’m not disagreeing but it was working fine before the layout change. Happy to try any recommendations tho!
-
I’m not disagreeing but it was working fine before the layout change. Happy to try any recommendations tho!
I don't understand.
Working fine before what layout change?
Did this amp ever work right?
-
There's 2 black wires hooked up to the 1st B+ filter cap ground lead. 1 of those wires go over to the solder lug strip with the -bias ground wire. Which I think now you moved that over to the preamp ground buss?
And the other black wire goes from that 1st B+ filter cap's ground lead over to the FWB's ground?
-
Yes it was sounding pretty good at one point, there’s a link to a demo of the reverb and trem in reply 15 but I tried to optimise the layout and rebuilt it and it’s gone to shit :BangHead:
-
It looks like both ends of your buss wire are connected to the chassis?
It can only be connected to chassis ground on 1 end, or you get a ground loop.
-
There's 2 black wires hooked up to the 1st B+ filter cap ground lead. 1 of those wires go over to the solder lug strip with the -bias ground wire. Which I think now you moved that over to the preamp ground buss?
And the other black wire goes from that 1st B+ filter cap's ground lead over to the FWB's ground?
.
Yes, that’s correct
-
Yes it was sounding pretty good at one point, there’s a link to a demo of the reverb and trem in reply 15 but I tried to optimise the layout and rebuilt it and it’s gone to ... :BangHead:
Ok, so it worked, then you changed some things and now problems.
So, what things did you change? How many things?
List all the changes 1 at a time please.
-
It looks like both ends of your buss wire are connected to the chassis?
It can only be connected to chassis ground on 1 end, or you get a ground loop.
No the PT end is insulated and it’s grounded at the input jack end
-
So, what things did you change? How many things?
List all the changes 1 at a time please.
I appreciate how ridiculous this sounds, I had some issues as explained at the start of the thread and with the help of some of the guys here I got it working pretty well however the engineer in me couldn’t resist trying to optimise the layout. I made some what I thought were relatively minor changes and wasn’t expecting such drastic consequences! I couldn’t list everything but the revised layouts are posted in previous reply’s, the schematic hasn’t really changed. I guess the question I’m asking is, it is making a horrible static noise now, what is that a symptom of? From reading online I am lead to believe it could be resistor noise hence focusing trouble shooting on the resistors although think I’m going down the wrong path and it’s more likely the new ‘optimised’ layout and grounding
-
Also wanted to post results just in case somebody thought about copying it!
-
Well, if it worked before, put it back the way it was.
A bad solder joint can cause all kinds of noise problems.
I thought you said it was oscillating?
-
Yes that might be the solution although I was hoping it might be a case of, amp sounds like a gieger counter, change component x. Seems not! Hadn’t appreciated what a fine art the layout was, ah well lesson learnt.
-
Yes it is also oscillating but was hoping to solve one problem at a time although starting to appreciate all these things are usually linked.
-
As far as the grounding.
The 1st B+ filter cap is the nosiest ground in the amp because of the charging pulses. This cap has the most current drawn from it and has the highest ripple in the B+ power supply chain.
So, you want to isolate these charging pulses from the rest of the power supply chain.
You have done part of it, you have the FWB ground going directly to this 1st filter caps ground lead. But, when you took the B+ power amp chassis ground from this point, everything that goes through that junction gets dirtied up by those pulses. You should pull the B+ power amp chassis ground connection from the screen cap's ground lead.
You want to try and get rid of as many chassis ground connections as you can.
Merlin shows using 1 chassis circuit ground. But, many guys use 2, 1 for the power amp, the other for the preamp. The PI gets grounded with the power amp IF it has a -FB loop. If it doesn't have a -FB loop, it gets grounded with the preamp.
And from looking at the pictures, you really don't have the circuits grounds tied to the B+ filter caps ground lead. You gather them together to form a ground star. So you make the caps star ground, send a single wire from the star to the buss, then form the next star and buss it over to the next star. This way the circuit's current draw can go round and round, without disturbing/modulating any other circuit.
This is why I said the -bias ground should be changed, -bias should be grounded with the power amps ground, which is where the chassis ground is for the screen filter caps ground lead. That is the power amp ground.
You need to read this, if you have read it, read it again.
https://www.valvewizard.co.uk/Grounding.html (https://www.valvewizard.co.uk/Grounding.html)
-
Yes that might be the solution although I was hoping it might be a case of, amp sounds like a gieger counter, change component x. Seems not!
Sometimes happens that way. It could be a bad solder joint. They can make the noise you seem to be describing. It can also be parasitic, wires in wrong place, grid wire to close to high output wire.
Hadn’t appreciated what a fine art the layout was, ah well lesson learnt.
Yep.
There are layout things you learn. You need a good grounding scheme, some space can help, wire lead dress, keep grid wires short and plate wires long if need be - after the coupling cap the plate wire is now a grid wire, shorten up excess wire runs, twist certain wires together tightly and evenly, keep output away from input, etc.
-
Yes it is also oscillating but was hoping to solve one problem at a time although starting to appreciate all these things are usually linked.
Wasn't doing these things before you changed some things, so they could be linked. :dontknow:
-
How was the oscillation stopped?
Is the rev2 layout in post 25 the one used? https://el34world.com/Forum/index.php?topic=32112.msg354945#msg354945
Chassis photo https://el34world.com/Forum/index.php?action=dlattach;topic=32112.0;attach=116474;image
Does the rustling noise stop if the valve in V3 is removed?
A full voltage survey may be beneficial.
-
I've not managed to stop the oscillation since the rebuild, with the channel volume at max, increasing treble or middle above 7/8 will induce squeal in the speaker. Unfortunately all of the anode bypass caps I installed to fix the first version are installed on this version so there is something else wrong. the 47pf cap across the PI plates isn't shown on the drawings or photos but I added it this morning and made no difference.
I tried playing a guitar through it earlier and maxing the reverb causes the volume to drop significantly so seems to be an issue there!
Also the tremolo seems to affecting the noise if it is enabled or not which is leading me to think it is an issue with the grounding.
The layout is pretty much as per the previous version I posted although I corrected a couple of errors, I've attached an as built and schematic. Photo is correct.
The rustling/static is significantly reduced when removing V3, lifting V3B plate or grounding V3B grid but not eliminated.
Yes I agree voltages are needed for further troubleshooting but it seems so fundamentally wrong atm that I think I'm going to change the grounding back to the previous version that I know was good (enough) and go from there, hoping to not have to do a full rebuild as I need to pay the family some attention :)
Plan is to restore the previous grounding and see if still bad.
If still bad replace all components/reflow all solder joints working back from the Phase Inverter.
If still bad revert to previous layout
I've got a couple of days off work next week so hopefully get some time to give it a good dose of looking at, will post results.
-
Strongly suspect this underboard wire to V3B grid is part of the problem, not sure how best to route it, might try a shielded cable
-
What's the yellow cap by the input sockets? I can't see it on the layout.
A screening plate around the input sockets may well cure the oscillation.
-
.... Unfortunately all of the anode bypass caps I installed to fix the first version are installed on this version so there is something else wrong.
Then the layout or something was wrong with this amp at the start, you put band aids on it.
-
Strongly suspect this underboard wire to V3B grid is part of the problem, not sure how best to route it, might try a shielded cable.
Why do suspect this under board wire?
Is it a grid wire?
-
.... Unfortunately all of the anode bypass caps I installed to fix the first version are installed on this version so there is something else wrong.
Then the layout or something was wrong with this amp at the start, you put band aids on it.
Yes that is why I rebuilt it! I was hoping the revised layout wouldn't need them but it's even worse! :BangHead:
-
Strongly suspect this underboard wire to V3B grid is part of the problem, not sure how best to route it, might try a shielded cable.
Why do suspect this under board wire?
Is it a grid wire?
Yes it goes to V3B grid.
-
What's the yellow cap by the input sockets? I can't see it on the layout.
A screening plate around the input sockets may well cure the oscillation.
.022 Tried it for RFI based on Merlin/Aiken but it made no difference, it's not connected now
-
I'm going over your latest schematic and lay out.
-
The 10K/.5w grid stoppers have to be soldered on the tube socket grid pin for them to work. With as little lead as possible with in reason.
on grid stoppers, how to use them;
https://www.valvewizard.co.uk/gridstopper.html (https://www.valvewizard.co.uk/gridstopper.html)
https://www.aikenamps.com/index.php/grid-resistors-why-are-they-used (https://www.aikenamps.com/index.php/grid-resistors-why-are-they-used)
-
You should read this also on grounding from Aiken Amps;
https://www.aikenamps.com/index.php/grounding
-
Your missing grounds in your trem on the layout.
There's a couple/few places that should be grounded back to the B node screen grid B+ filter cap ground lead.
-
Not sure which ones are missing :w2:
I updated the trem grounding earlier, looks like this now. Not made any difference, will look at reverb next.
-
Not sure which ones are missing :w2:
That's a problem.
This is where you go back and fix your layout drawing.
You need to fix your layout and then match your layout to your actual amp.
-
Ok there is a link missing across the bias grounds on the layout. It is installed on the board although I just chopped it half to make the changes. The trem grounds where all on the layout, I’ve just moved them tho
-
You have to up date both your schematic and layout when you make changes so you can find your way and so we can see if we see any things wrong.
-
Will update it when I get chance, wife and kids getting in the way of my amp building :laugh:
-
Will update it when I get chance, wife and kids getting in the way of my amp building :laugh:
Amps come and go, love on the family. :icon_biggrin:
-
Did you ever try disconnecting the -FB loop?
Disconnect it either at the output speaker jack or disconnect the -FB R, 820 ohm, R51.
Any better?
Reversed OT primary leads with a -FB loop doesn't always cause a high pitched sequel, it can cause all kinds of weird things.
-
Yes I should have written this down but was one of the first things I tried as suspected the wire routing, disconnected but no discernible difference so it’s back in the circuit.
-
I'm measuring around 0.7 - 1 VDC downstream of V1B coupling cap C13. I understand there should be no DC present here, is that correct?
I've changed C13 for a new cap but the DC is still there. It peaks at about 9VDC when I first flip the standby switch then decays to around 1V. Not sure where this DC is coming from??
-
I'm measuring around 0.7 - 1 VDC downstream of V1B coupling cap C13. I understand there should be no DC present here, is that correct?
I've changed C13 for a new cap but the DC is still there. It peaks at about 9VDC when I first flip the standby switch then decays to around 1V. Not sure where this DC is coming from??
Correct there should be 0 V DC.
With amp off and deenergised, what's the resistance to the chassis at that point?
-
3.47M
-
Suspected there was some leakage on the board earlier, it was looking pretty dirty after all the mods so I gave it a good cleaning with isopropyl and some contact cleaner, it’s down to 150mv dc there now. Still scratchy static noise but reduced
-
Think it's back to the drawing board with this one, I'm going to move some things around and rebuild it. Thanks for everyone's help so far, got some good info to work with. Will report back when complete although it might take me a while.
Cheers
-
Did you try replacing C17?
Greg
-
3.47M
That's too high.
Edit - oops it's fine.
-
Did you try replacing C17?
Greg
Hi Greg, yes I had already replaced C17. Just feel I’m chasing my tail with this now, may stumble across the static/rushing water noise issue by replacing a component or reflowing a joint eventually but the feedback issues are still going to be there. I suspect the tone stack is too close to the reverb drive circuit. I think I can rearrange that section by adding half a dozen terminals but obviously need to remove the board.
-
something like this
-
Did you try replacing C17?
Greg
Hi Greg, yes I had already replaced C17. ...
If a leg of C17 is lifted and the amp powered up, is there still voltage on that turret / node?
-
I suspect the tone stack is too close to the reverb drive circuit. I think I can rearrange that section by adding half a dozen terminals but obviously need to remove the board.
I was going to suggest this for your rebuild. It greatly shortens up the grid wire feeding the TS. And it makes room on the turret board. Now you can and should move the B+ D node filter cap over to where the TS components were. Grounding is now much better too.
No, you can hang most of the TS caps and R on the TS pots. You would only need a couple turret's to pick up the other side of those components.
-
And don't put any grid wires under the board on the new layout.
-
Board updated
-
Board updated
And?????
-
And.....
it's still under construction!
Got a busy weekend (actually going to 2 gigs!) so won't get it finished until next week, I know you're all as keen as I am to see if it works :laugh: Here's a another progress shot in the mean time.
-
OK so I've finished rebuilding it and appears I've made some limited progress, the oscillation when the controls are maxed has stopped although I still have the rushing water/static noise. The voltages are as follows;
V1
Pin1: 187 Pin2: 0 Pin3: 1.73 Pin6: 187 Pin7: 0 Pin8: 1.75
V2
Pin1: 333 Pin2: 25mV Pin3: 4.82 Pin6: 333 Pin7: 25mV Pin8: 4.8
V3
Pin1: 194 Pin2: 0 Pin3: 1.62 Pin6: 195 Pin7: 240mV 0 Pin8: 1.61
V4
Pin1: 178 Pin2: 52 Pin3: 82 Pin6: 199 Pin7: 53 Pin8: 82
V5
Pin3: 440 Pin4: 438 Pin5: -41 Pin8: 17mV
V6
Pin3: 440 Pin4: 438 Pin5: -41 Pin8: 20mV
It looks to me like the noise is resulting from the 250mV DC on V3B grid, I measured 0V DC here on the initial build when the amp was working successfully.
I've tried lifting one leg of C17 as suggested by pdf64 but this makes no difference
I have observed that the voltage rises to around 500mV as I increase the reverb pot to maximum.
The voltage is still present if I remove V3.
I can't see where this voltage could be coming from other than a leaking coupling cap from V1B (C13) however I have tried swapping this cap previously with no change. I don't have any spares so will have to order a replacement, does anyone have any suggestions in the mean time?
I suspected the board but I thoroughly cleaned it with isopropyl when I rebuilt that section and I'm not measuring any dc voltage on the board around the terminals.
-
...
It looks to me like the noise is resulting from the 250mV DC on V3B grid, I measured 0V DC here on the initial build when the amp was working successfully.
I've tried lifting one leg of C17 as suggested by pdf64 but this makes no difference
I have observed that the voltage rises to around 500mV as I increase the reverb pot to maximum.
The voltage is still present if I remove V3.
...
How about if the valves in V1 and V2 are removed?
If that doesn't help, how about if C13 (3n3F reverb recovery output) is lifted?
-
Just tried both suggestions and no change,
The only thing that stops the static noise is grounding V3B grid, I am totally stumped here :help:
DC resistance to ground from Pin7 is around 150K depending on the Reverb pot setting, I can't see how it can be getting a DC voltage.
I swapped V3 tube socket when I did the rebuild in case that was the fault, also tried several 12AX7s there as well but no change.
-
Sorry misinterpreted your second suggestion there regarding the reverb recovery coupling cap, C13 is the V1B coupling cap, anyway I've just also lifted C19 the 3.3nF cap and no change.
So for clarity I've lifted C13, C17 and C19 and there is still 250mVDC present on Pin 7 in all three scenarios.
Only grounding Pin7 eliminates the static.
-
Forgive me, I haven't checked through the full thread, but do you have a grid leak on that triode (220k typical) or did go forgo that as in the Princeton Reverb?
If you disconnect the wire to the grid, do you still measure a voltage at the socket? If so, and since you've ruled out tubes by swapping a few in, then you may need to look at the socket itself.
If you disconnect the grid wire and don't measure DC at the socket but do at the board then there's an off chance the black pigment in the fiberglass is causing it to be leaky. This is provided there are no wiring errors or alternate paths for DC--so be sure you've ruled everything else out first. Black in plastics (resins included) is typically derived from a carbon of some sort. It even affects the mechanical properties of plastics compared to other colors. It looks cool compared to the sickly standard color, but I'd be hesitant to use it in high voltage applications. Not because of a track record of leakiness, but because of the increased possibility of it and the availability of alternatives without that risk.
-
Yes it has a 220K grid leak (R38 on the schematic) which has continuity to ground.
If I disconnect the grid wire the static noise goes and I measured around 100mv on pin 7 which decayed to around 6mv over roughly 2 minutes. There is a loud hum while the meter is connected.
I measured 78mV at the terminal on the board.
You will have to forgive me but I honestly have no idea how to interpret this!! The colour of the fibre glass is however one of the things that has changed since it was working, you can see from the first pictures it was the standard green colour originally.
-
You will have to forgive me but I honestly have no idea how to interpret this!!
There are only two options left.
- The coupling cap is leaky in a way that has eluded your efforts to identify it.
- The board itself is conductive enough to allow DC build up at the turret.
Try "floating" the connections to the grid instead of terminating everything at the turret/eyelet. It's going to be messy and you'll have stuff sticking up in the air, but it's just for a temporary test.
If this fixes it, then I'd think the 2nd option is very likely. The only way to confirm for sure is to use a "Megger" to measure the insulative integrity of the board.
-
Don't think I'll be able to get my hands on a Megger any time soon!
In hindsight there's not much clearance between some of the terminals, I've just doused the board a couple more times with isopropyl to try and get rid of any remaining residue, will let it dry out overnight and try again in the morning but not holding out much hope as it already looks pretty clean.
I'm still clinging on to the fact that it has worked previously and trying to think what could be different to before. I was also thinking about floating the connections but not sure how feasible/safe it would be.
Might have to be another redesign on a green board with extra spacing :sad2:
-
I'm still clinging on to the fact that it has worked previously and trying to think what could be different to before.
Yes, but it only worked with band aid by-pass caps. It always had problems. :w2:
-
Yes, but it only worked with band aid by-pass caps. It always had problems. :w2:
was referring to the DC voltage on V3B grid that wasn’t there originally.
-
I've done a bit more trouble shooting this morning and can only conclude that there is DC leakage on the board.
I've disconnected C13 and C19 and measure around 500mV at the junction of R26/C17, I am also measuring this on what is now a spare terminal (out of circuit) alongside this junction. If I clip a lead from the disconnected side of C13 directly to V3 pin 7 the static noise goes and the grid voltage drops to a few millivolts.
What I was struggling to understand was if there is leakage via the board why is this only manifesting as an issue on this specific grid, i.e. surely I should see this on V3A then it occurred to me that V1 and V3A are connected to the pots/jacks directly so no connection via the board.
There are a few spare terminals on the board now as a result of the changes and I'm measuring various DC voltages to ground on all of them.
I obviously can't say if it's the board itself or something I have done to it but looks like the only solution is a rebuild on a new board.
Thanks for everyone who has taken an interest so far, really appreciate the help!
-
Oh dear, that's unfortunate!
What material is the board, where did you buy it?
It looks a fibreglass type, so I thought this issue would be very unlikely, though it seems more common with the vulcanised compressed paper fibreboard traditionally used in vintage Fenders.
So, well done tracking the root cause issue down.
-
Yes it seems so unlikely that I had pretty much discounted it but I can't see any other explanation.
It's G10 fibreglass board bought from a reputable supplier here in the UK, I'm reluctant to say where because I may well be completely wrong given my limited diagnosis skills! I will reach out to them to see if they have had any other issues.
I have obviously done a lot of rework and whilst I have tried to clean the board the leakage may well be a result of my work somewhere!
-
Well I'm very pleased to report that the supplier of the original board sent me a green replacement and after a rebuild the amp is working again, static oscillations and hum all gone!
I admit I've left all of the 'band aid' caps in place as I was working on the principle that it's much easier to install them during the build and remove them later but I'm going to enjoy the amp for a bit before I change anything, going to give the soldering iron a rest for a bit :icon_biggrin:
-
Picture of the now working amp
-
Well done for sticking with it, that was a weird bit of bad luck.