Vox AC100/2 build

[SoundmasterG]

That amp has too much overdrive/distortion and a lot of those weird sounds you are getting are likely oscillations. A stock JMI circuit Vox AC100 is a very clean amp and will only start to go into overdrive when it is all the way up. The rhythm guitar and lead guitar (and bass) parts from the Beatles "I Feel Fine" were done with a Vox AC100, so you can hear Lennon's rhythm part. That is with the amp all the way up....that is all the overdrive/distortion it should put out. That is way less than your recording showed.


What I would do is hook the amp up to a resistor dummy load, hook it up to a signal generator outputting about 100 mV, and hook up your oscilloscope and check the various stages to see how the signal goes through the amp and how it distorts at each stage. Despite what I said above regarding overdrive/distortion levels, you will see overdrive and distortion on the scope, but it will sound mostly clean when you listen to it. You should be able to ascertain the function of all the controls on the amp with the scope and sig gen before you even plug it in to play it. Get it distorting at the output of the amp and then back off a bit and check each stage and see if one distorts before the other., etc. Check the voltages at each stage also. Sometimes you can have too much gain from one stage and it will overdrive the next badly causing grid current distortion and blocking. You can also get oscillations due to poor layout or poor soldering or a lack of shielding in certain places. That can also be caused by bad grounding. I forget if you started with a stock AC100 circuit, or if you modified it for more gain, but I would suggest to go with the stock circuit, get that working, and then modify it from there. The use of the 12AU7's where they used them in the Vox is key to that sound also. If you use 12AX7's instead, then it doesn't get that same harmonically rich sound and is too overdriven.

Greg

[pompeiisneaks]

Greg, thanks, I've got a from scratch build of a Vox AC100/2 based on the chassis of a Vox AC100cph.  Therefore it is a bit of a mild frankenstein.  I'll give that a go.  I have been playing a bit with the scope to see where it seems to go berserk, but I'll try it as you've said and see if I can find where the distortion is originating.  I've got the right tubes in now.  The 12AU7 at V1 and V4 per my schematic. I may end up completely removing the reverb circuit and seeing if everything else gets better, and if so, try and figure out why the tried and true 1 tube reverb is the issue (it may not be but I'll see).

~Phil

[pompeiisneaks]

Okay so tonight I replaced that resistor that I blew, and also returned the two 100 Ohm to ground on the heater line.  The hum seems a lot quieter now, with it biased a bit better.  I realized, though, that the bias is a bit too hot.  The current is only about 20mV.  The tubes expect about 40mV at 70% dissipation, if I'm correct. 

Here comes the 'something is wrong' moment.  I slowly bring up the bias pot until I get near 40 and about 36 or 37mV it starts humming loudly and the voltage, without me touching it, starts climbing to about 44 or so and I hear that click/pop and I've got the volume down to 0 on the speaker.  after that the mv drops to something like 0 or even -5mV and then climbs back up to about 36 or so.  If I dial it back just a tad, to about 34, it doesn't seem to do it.  I'm guessing it will if I play because that will cause the current to climb (its not class A so it will right?)

Does that mean somethings being overdone in the tubes themselves?  I'm wondering if this circuit needs lower voltages?  The schematic for the original states the power tubes should be about 450, and mine are about 480, the schematic also calls for much lower voltages on the preamp tubes than what I'm getting, I'll copy/paste from my previous measurements on this thread.  Am I just running at a too high of a range for the tubes?

One other thing I'm thinking could also be it.  The voltages for the reverb are really hot too, like 400 ish, and I'm guessing that circuit, with the preamp tubes, would do better with something like 150 to 200 V?  no? 

Do I need to adjust my voltages across the board with a dropping resistor?

Here's what I get on my A-E points:

A: 504V, B: 484V, C: 481V C': 476V, D: 320V, E: 293V


but the amp asks for:

A: 470 B: 450, C: 430, D: 310, E: 270

That means the currently setup 22.5k resistor in R23 (the big 25 Watter Steve found was way too lacking). 
may need to be higher to bring voltages down to what's expected.  Or do I want to put one inline before A to drop A down to 470 and the rest should just fall in line? 

Do I need to put the reverb connections from B and D to something a bit lower?  (after adjustments mentioned above).  so that they're a lot cooler for that reverb transformer and the circuit in general?

I'm still confused about how I can drop the voltage.  I know I need to put in a resistor in series before the A point and that will help drop the Voltage, but I think you need to know the current, is that the sum of currents of the plates for all tubes? 

ugh, still a noob at power supply design!

I've been playing with PSU designer but I don't know how to setup a bridged diode like mine has (two of them) so its also odd in that way. 

Any input on what's going on?  Or are my voltages all 30V too high nothing to worry about?

~Phil

[SoundmasterG]

I would disconnect the reverb circuit for now and get the stock amp working first. Reverb circuits are notorious for causing hum and noise if they are not correct. Go to the exact original circuit and get that working. After you have the amp stable then you can start adding changes one at a time.


I don't recall what the voltages were in my build and my notes aren't handy right now so I can't help you too much with that, although your voltages do seem a little high, though they should work fine as long as you have the bias set correctly. The AC100/2 schematic calls for raw B+ of 470 V and 466 V at the tubes themselves so you aren't that far off. What are the voltages at the tubes? Pins 1 and 6 are the plates and pins 3 and 8 are the cathodes for the preamp tubes. There is the supply voltage to be concerned with, and also the actual voltage that the tube is seeing. The power tube plates and screens matter, as does the preamp voltages. The way to control those voltages are the dropping resistors which are between the nodes and caps in the power supply. Changing one early in the string such as the ones between A and B will change all the voltages downstream also. The exact voltages aren't super critical, but they are important as they determine how much headroom each stage has.


Before you start in with any adjustment to the voltages downstream though, get the bias set correctly so the amp will be stable. When you talk about current, you should be talking about amps, yet you are mentioning mV. I am assuming you are using EL34's? Those are a 25 watt dissipation tube. 70% of 25 watts is 17.5 watts and 60% is 15 watts. The bias should be set somewhere in that range for the amp to be happy. So if the B+ is 504V, then around 34 mA at that voltage gets you close to 17.5 watts, and 29.8 mA gets you close to 15 watts. (504 * .034 = 17.136, 504 * .0298 = 15.0192) The B+ will change however as you adjust the bias up and down, so you have to remeasure as you are making each adjustment and do the calculation again. (higher B+ with lower bias current, conversely lower B+ with higher bias current) You can figure out the bias current if you use a Bias Rite, or a 1 ohm resistor from pin 8 to ground and measure voltage across the resistor. Since it is 1 ohm then whatever current is going through the resistor will convert straight over to the voltage that you measure. Notice I haven't mentioned the bias voltage yet....that will be whatever is necessary to get the tube idling at the current you want...that is all the bias voltage matters...the value of it isn't important otherwise...the current through the tube is what matters. Keep in mind that that the cathode bias current also includes the screen current in addition to the plate. (which is what you are measuring with the 1 ohm resistor to ground from the cathode of the power tube) The error due to the addition of the screen grid current is in a safe direction so if you set it for 70% of max dissipation, it is actually a little less than that.

Greg

[SoundmasterG]

I know it is frustrating with the problems you are encountering, and you want to play the amp, but get the amp back to stock, and get the basics right first and then start to tackle the problems ones at a time in the method I mentioned previously in the last post. It is the quickest way to solve all the issues. You can't begin to troubleshoot anything until the amp will sit there for hours without complaint while you troubleshoot the live circuit. A lot of the weird pops and distortions can be due to oscillations, bad grounding, and also bad soldering or parts accidentally touching if you didn't do the layout very well so it is worth it to check everything with a fine tooth comb if you still have issues after getting back to stock.I did my AC 100 build in a Sovtek Mig 100U chassis. I used the Sovtek power transformer and used a Heyboer OT and choke. The voltages were higher in mine than stock also...I think about 470 V B+ instead of 450 V. I was aiming for a combination of the early cathode biased ones and the later fixed bias ones so my preamp voltages were tweaked to be in the middle of both. The early cathode biased ones ran VERY hot in a small chassis and liked to burst into flames sometimes. The later fixed bias ones still run hot though. Mine is using switchable cathode bias and fixed bias though if I had to do it again I would just stick with the fixed bias as they sound very similar to each other and the amp runs VERY hot when in cathode bias mode. I had mine biased about 65% of max dissipation when in fixed bias and it was still running quite hot compared to the typical Fender design. I also kept blowing screen resistors until I got them sized large enough. ( I think 7.5 watt is what are in there now.) I ran into an issue with mine with the layout and my second channel I added and was going to have to gut it and start over and I haven't done that yet so mine isn't finished but it is stable aside from that oscillation when I get up to about 8 on the higher gain channel, and the stock AC100 channel works perfectly and sounds great and is also clean until it is almost all the way up. That is what yours should do too in a stock configuration.If your voltages are climbing like that then you have the bias current too high most likely (bias voltage injected into the grid too low). The amp will hum a lot when the bias is incorrect. You should be setting the bias when all controls are at zero and you are not playing the amp. It should also be going into a resistive load. Get the bias correct so you don't blow up any more tubes and then start addressing the other issues after that.


Greg

[pompeiisneaks]

Thanks for the info.  I'm mentioning mV because I've got 1ohm resistors on pins 1/8.  I should have made that clear.  This means I was near 70% for the tube of 40mv or 40mA which at 480v is 19.2 watts, or 70% (or near to it).  That is when things start climbing like mad and getting near 50 before the pop and then dropping to near 0 before climbing again.  I definitely think I will remove the reverb circuit to see what that provides, I think it may be part or all of the problem, then I need to figure out why it is so off. 

~Phil

[SoundmasterG]

Ok, so you are reading the current as mV but it is really mA that you are indirectly reading. Good that is clear, thanks.


If I do the math, 70% of 25 watts dissipation is 17.5 watts and 60% is 15 watts. 19.2 watts is more like 76.8% of the max dissipation for an EL34....these amps run hot so you should be running it down around the mid 60's rather than the mid 70's, especially if you want it to be stable so you can do testing. It doesn't hurt things to run it at 60% of the max if it allows the amp to behave. You can get everything right and then turn the bias up later if you want to experiment but I'd advise to cool the bias off. The reason things start climbing is because the tubes are likely going into thermal runaway. Are you sure you aren't getting orange or red plating? You could also have a socket that has arced and if that is the case nothing you do will work until you replace the socket. What kind of sockets did you use? If you aren't careful you will blow out a set of power tubes. As I said before, once they have a bias failure and get too hot, they are never the same and are susceptible to thermal runaway again.


I once had a Bassman 100 that was fine on the bench but would blow a fuse and a set of tubes when played live. The problem ended up being the socket, but it took three times for the problem to show itself on the bench, and when it finally did, it behaved the same as what you describe...it would start to bias creep and then go out of control.


Greg

[sluckey]

A: 504V, B: 484V, C: 481V C': 476V, D: 320V, E: 293V

That is a little high but EL34s work well with those voltages. What voltage are your filter caps rated for?

With the issues you've had, especially the plate to heater short in the EL34, it's possible that the OT has been damaged from over current. Let's hope not, but keep that thought in your head.

Pull two of the EL34s and see if you can adjust the bias current up to about 35-40mA and keep it stable. You must still have one tube on each side of the OT primary, so don't pull the wrong tubes.

[pompeiisneaks]

The caps are all either 350 or 450 and are in quads of parallel and series in the A and B stages, the rest are 475V 16uF FT's (the amp had 6x 100uF 450V and I bought two more from doug that were 100uF 350V (which means that the max power on that quad would be 700V instead of 900V but since I'm at 484 on the B point, I'm good there too).

I'll try pulling the two tubes. 

Greg,

I've got an app that is called Tube Bias Calculator that says for an EL34 70% dissipation for the tube is 36.5 mA for a AB fixed bias.  I was trying to get it near that ballpark without going over, and that is when it started taking off.  The tube sockets are brand new, so I'm hoping they aren't arced.  I did a test of the pins to make sure they weren't shorted out to any other pin at one point in this long, long thread :)

~Phil

[pompeiisneaks]

Haven't had a ton of time yet today, but during a short work break, (WFH today) I went down to the garage, and pulled two tubes and did a bit of biasing.

One thing I realized was that my assumption (made an ass outta me and umption) of 1ohm was wrong.  The resistor on the one I was biasing from is .8 ohms.  Which makes it a lot hotter than I planned. 

At any rate with two in any of the sockets I can't seem to get any problems at all.  When I put all 4 back in, they definitely start going off again.  I did make the realization of the above at that point, and realized I'm at 85% dissipation /doh. I dialed it back to about 30mV on that one, and it puts me more like 68% and 35.4 mA. 

Basically, I should have been wary of that and forgot.  Also I have others that sit more like 1.1. ohms as well.  I think the bias was too hot at my '35-40' mV range and now its much better.  I'll also want to play around with removing the reverb circuit, but I may very well have a better sounding amp after just that.   

I'll report back a bit later if I can get some more down time working to monkey with it a bit.  I think time to mess with the reverb circuit removal may have to wait until the weekend.

~Phil

[pompeiisneaks]

Okay, here's something quite odd.  I've just put the scope on, and see this.  I'm attaching a screenshot of the scope. 

The yellow trace is my input connected directly to the jack. The blue is at the 47k resistors right before it hits the power tubes. 

The bigger deal here is that when I turn the volume up a bit, the INPUT gets all kinds of odd distortion....  is that a ground hum coming into the jack itself?

Let me know what that kind of noise means?

Edit: the input is normally very clean until I turn up the volume a tad.

~Phil

[shooter]

I'm sorta lost in the weeds on this build, but if I seen that at my input, i'd scope my input signal 1st, NOT plugged into the amp, if that's good, i'd pull all tubes except V1, then put trace 1 on the input jack, trace 2 on the output of V1A, AFTER the coupling cap.

[pompeiisneaks]

Maybe I'm not explaining it very well.  (I have done what you said and it is compeletely clean there on the signal input not connected to the amp).

The signal is super clean if the amp is off or at 0 volume.  Once I up the volume a little, that noise shows up on the input connected probe. 

I don't need to go further than that. 

~Phil 

[pompeiisneaks]

Oh I just reread that and now I get what you mean, figure out if the interference is gone with only V1, then add V2, etc until it shows up, and then figure out what stage is the cause

I'll give that a try.

~Phil

[shooter]

 

figure out what stage is the cause

circle gets the square!

[pompeiisneaks]

Well just tried that and found 'where' it is happening, V1 and V2 (input and tone stack) have no issue, as soon as I added V4 (skipped V3 related to reverb) it came back.  I'm guessing, though, that the issue may be in the reverb part of the circuit.  I'll take some time tomorrow to desolder the parts related to the reverb circuit and jumper past the 100k resistor between the tone stack and the PI that was where the reverb circuit lives so it goes strait to the end... I'll see what's up.  I've got to go see Deadpool with my daughters tonight :)

~Phil

[SoundmasterG]

Greg,

I've got an app that is called Tube Bias Calculator that says for an EL34 70% dissipation for the tube is 36.5 mA for a AB fixed bias.  I was trying to get it near that ballpark without going over, and that is when it started taking off.  The tube sockets are brand new, so I'm hoping they aren't arced.  I did a test of the pins to make sure they weren't shorted out to any other pin at one point in this long, long thread :)

~Phil

Phil,


You should stop using that Tube Bias Calculator and just figure out how to actually do the math yourself. For a 25 watt tube like an EL34, you multiply 25 times 0.7 for 70% and times 0.6 for 60% and that tells you the range you have to work with. If you were using 6L6GC's, then since those are a 30 watt tube you multiply 30 times 0.6 or times 0.7 and you can figure out your safe bias range with that one too. Then you measure your plate voltage and multiply by what you read across your 1 ohm resistor. (say you measure 37 mV and 480 V, so you take 480 times .037 which = 17.76 watts) Its that easy, but remember that every time you adjust your bias voltage, then plate voltage will change as will the bias current so you have to re-measure and recalculate, and different tubes can pull different current from each other too, even when paired with each other. If you have them too far out from each other with their current draw, then the amp can hum excessively.

For the sockets, look closely at them and see if you see any evidence of arcing. Ceramic sockets are immune to arcing, but the plastic types and micalex etc., can and will arc sometimes. There is something going on....either a arced socket or a bad tube in your quad that causes the amp to misbehave when you try to get stable bias.


The scope shot you posted is showing high frequency oscillations at the top of the waveform there, but they seem to be there from your signal generator source unless you have that scope shot somehow of something else? As a comparison, on my AC100 build on the clean Vox channel rather than the high gain channel that I added, I could turn the amp almost all the way up with the signal coming in from the signal generator and see the signal at the power tubes much larger but still a clean sine wave. As I turned it all the way up then I just started to get some overdrive and distorting of the waveform, but no oscillations like that.


One big problem with your amp that may be causing some of these issues is the layout. The sockets for the preamp are over on the left and your board is over on the right and every sensitive grid wire (grid wires are like antennas) from there has to run over to the sockets in parallel with other stages and also with plate wires, cathode wires, and whatever else you have in there....that is a good recipe for oscillations and weird noises. Take a look at a typical Fender layout....the input comes in, goes across the board, and then to the sockets which are right behind the board. Then it comes back up to the board and goes up to the tone controls, back to the board, back to the sockets etc. The stage wires don't cross each other, or go parallel with each other in long wire runs like what you have going on. The Fender layouts aren't perfect but they are good and work with most amps for the most part.


Greg

[pompeiisneaks]

Greg,

Thanks,  Sadly the chassis is as it was for the old Vox AC100CPH, so I don't have the ease of switching it around much, I do understand what you're saying, though.  after grounding the heaters the hum has gone down significantly.

As for the bias, I already knew that method, and have used it in the past, but I like calculators that do the math for me.  Ultimately, though, if you see one of my last posts, it was due to my 1 ohm resistor actually being .8 ohms that I was so far off.  I promise, the input signal is clean.  really clean.  If I have the amp off, or remove the cord and plug directly into it, there is a perfect sine wave.  I just showed the shot of what happens as soon as I turn the volume up from 0 to about 2 of 10 and that shows up immediately.  Per the troubleshooting I did yesterday, I found that it goes away until I put the PI Tube back in, but I suspect its the reverb part of the circuit causing that.  the popping is the tubes going into runaway due to me biasing them at 85% or so because of estimating wrong with that .8 ohm resistor.  I've dialed that back so that its near 65% I think and will get the reverb issue resolved if not removed and move on from there.

~Phil

[pompeiisneaks]

Oh and the power tube sockets are ceramic, so per what you mentioned, they should be solid.

~Phil

[pompeiisneaks]

Okay I've pulled the reverb part out completely, and have been doing some testing.  I've found a few things. 

1. it is still distorting quite early. 
2. It is still getting into runaway mode even at hotter bias settings. 
3. I did some voltage measurements with the oscilloscope at the 4x47k resistors that split out to the 4 tubes as well as some bias measurements on either side of the tubes and found that they're off quite a bit.  Example:

left half of PP about 34mV and right half about 9mV
left half of output from PI at a fairly low volume 62V right half at same volume, 40V

4. The output voltages from the PI PtP seem pretty high.  With the sine wave input going into it, and with volume up to max, I was getting about 500V input.  (I kept thinking I had my 10x probe set to 1x or something, but it wasn't, on the scope it was set to 10x and on the probe it was at 10x.  That seems pretty hot for input into a EL34 no?  Also I'm seeing plate voltages climb to something like 100mV when I get the volume up there so I turned it back down pretty quickly.  (maybe that's due to it being a sine wave?)

I see two potential things that need sorting. 

1.  Fix that imbalance, I'm not sure why its so off, but I measured all the resistors in the region of that circuit and they all seem right, the ground ones are at 1M the R19/R20 level, R21/22 add up to like 174k in circuit (2x330 is about right, a touch high but seems fine)
2. figure out why it seems to be so hot coming out of the PI. 

So in looking at the EL34 datasheet, it seems to imply that it should be 50V RMS input to the grid, thus my worries that at even nominal volumes input its so high... what would be causing that?  Is there something wrong with the bias on one or several of the preamp tubes pushing it that high? 

I'm going to do some re-measurements to ensure I'm not going insane, but I definitely saw the imbalance. 

~Phil

[pompeiisneaks]

One other thing to note, pulling the reverb didn't seem to clean up that odd static, but then I kept wondering why my probe was so dirty compared to the other (the blue one on the screenshot).  It ended up being messed up, I couldn't even get a clean line on it on the on scope reference square wave, so I switched them and the noise seems to have mostly gone away.  I also resolderd the connections to my speaker and that also seems to have removed some of the worst of the crackling.

As for the hum, its only really bad if I've got my guitar close to it and when I walk about 5 feet away the hum is pretty low and at 10 feet it's almost gone, so I think its just noise from the p90's.

~Phil

[pompeiisneaks]

Nvm on that odd static, it's back.  I did more testing, probing etc,  Here's the latest voltages of all tubes and HT rails:

Preamp, Rows down are V1-V4, columns across are Pins 1, 2, 3, 6, 7, 8 (note V1 and V3 have no second half in use, its the reverb for V1, V3 first half is reverb return, but nothing is coming into that tube right now)
   

88.7V	69mV	2.8V	-	-	-
196.6V	28.8mV	1.5V	323.2V	198.7V	199.2V
319.9V	95mV	79mV	-	-	-
278.5V	310mV	11.9V	280V	159mV	11.9V

Power, rows down are V5-V8, columns across are Pins : 1/8,3,4/6,5

26.8mV	477.5V	463.8V	-38.90mV
29.5mV	478.9V	464.9V	-38.93mV
20.7mV	482.6V	469.4V	-42.19mV
21.0mV	482.3V	469.2V	-42.01mV

The Power rail (HT/B+

A: 487.7V, B: 467.9V, C: 468.9V, C': 450.0V, D: 321.9V, E: 279.5V

I'm still boggled...

Like I said, when I did test with the tubes out that 'static'/noise disappeared until I got the PI in, but I did have the power amp tubes in, do I want all of them out so I see if the PI isn't really the cause?  (I though I was just pulling and replacing the preamp tubes to see.)

~Phil

[SoundmasterG]

Good the power tube sockets are ceramic....that should rule out arcing at the socket. Since you had the bias way too hot previously, some of the tubes in the quad may be defective now. You could try to run just a pair as you did before but swap tubes around so you can see if one or more in the quad is having issues...but before you do that, try this.....take the power tubes out. Leave the preamp tubes in. Run a clean sine wave about 100 mV into the input of the amp. Trace the signal through each stage at various volume levels and see if you have any issues. You should be able to measure the clean signal on the grid, and then measure it on the plate and see a much larger signal that is inverted and without weird anomalies at each stage. You should be able to see the effect when you adjust the tone controls on the signal if you measure after the tone controls with the scope probe. When you get to the phase inverter, you can measure the voltage output when the rest of the preamp is a clean sine wave and see what you get. The voltage output there will be a bit less when you put the load of the power tubes back in, but this will give you an idea if the preamp of the amp is ok or not. If not then you can fix until you have that sorted and then move on to adding the power tubes. As Sluckey said before, divide and conquer. Simplify the situation and get one section sorted and then move on to the next. Removing the reverb is good even if it is not the cause of the issues as it simplifies what you are trying to do to find the problems.

With the layout, it looked to me like you should be able to put a small board in between the preamp sockets and the power tube sockets...but if you can make it work as it is now it might be easier than changing it up. I would try to run your wires between the board and socket a bit more organized so you don't have different stages running between the sockets and the board right next to each other, and perhaps use shielded cable for the grid connections. I would also think about adding grid stopper resistors right at the sockets as Merlin mentions on his site and in his book. That may help if you are having trouble getting it to behave.

I forget about how close the two sides are in that phase inverter design. It is a floating paraphase circuit....you can see that on Merlin's site. It doesn't balance perfectly, but it should balance pretty well. You're still using the 12AU7 in the PI slot right?

Greg

[pompeiisneaks]

I am using the 12AU7's in the V1 and V4 PI stages, yes.

I just did some more troubleshooting, and I'm suspicious the issue is with the OT.  I had the power tubes out, and can't get the weird issue to happen at all.  I put in Two tubes, and ti works fine as well on the scope,

As soon as I put in all 4 tubes it starts freaking out.  Additionally I tried swapping tubes, not only in and out only two at a time, but also the socket the tubes were in (keeping them to their own half of course) so that if I had left side tube 1 and 2, first time I did 1,2 then 2,1 and with just 0 1, 1,0, 0,2, 2,0 for each case (i.e. 0 meant socket empty) and mirrored the same combo on the right side as well, I never once got the issue with just two tubes in, but eveyr way I try with 4 it does that stupid noise issue. 

I am also suspicious per something that sluckey said about the OT, so I tested the impedance of each half, and I get 19.8 ohms on one side and 16.7 on the other.   I checked and found someone selling them and lo and behold I see this:

http://www.etronic-parts.com/product_info.php/info/p1090_VOX--Output-Transformer-AC100CPH.html

they're supposed to be 18 ohms a side, is a variance of 2 ohms enough to matter?  Is the OT what's wrong here?

~Phil

[SoundmasterG]

I took a look at my notes from my build. B+ on mine was 467 V. Power tube plates were 460 V and power tube screens were 452 V. The negative voltage going into the power tube grids for the bias was -35.9 V. I don't know what the mA were that the power tubes were biased at as I didn't write it down on these notes. The phase inverter was all mucked up as I had been playing around with different circuits there and so I don't think the voltages that I had on these notes are going to help at all and I can't find my notes from when I had the stock PI circuit. I had a 440 V supply to the PI, and 352.3 V, 344.4 V, and 328.8 V to the other nodes. My nodes were also supplying the high gain stage in my amp though so they would be different than a stock circuit, plus I had adjusted dropping resistor values to get the supply a happy medium between the fixed bias and cathode bias versions of this amp so my values are all different than stock. The first gain stage in my amp had 91.5 V on the plate and 3.25 V on the cathode. This is with the stock 12AU7. The second gain stage which is the 12AX7 is at 214 plate and 1.58 cathode. Third stage is the cathode follower and it is at 352.3 V plate, 214.3 V cathode.


I also attached a picture of mine so you can compare. Keep in mind that mine was a work in progress that will have to be gutted and started over. My board was split in two due to the Z mount of the OT. I got some attachments for the OT so I can mount it above the chassis and will be doing so when I get back to this project. This will allow a full length board, which will put the circuitry for the PI right next to the socket instead of several inches away on the second board. I was getting oscillations on my high gain channel at the PI due to the long runs to the circuitry from the PI socket, and the fact it was passing right under an unshielded OT. I also had some experiments going on and had tacked some parts in so it is sloppy right now around the PI and the power tube screen grid resistors too. I also had paralleled film caps with the e-caps in the power supply to check that out. I won't be adding that to the next version so that will simplify some things. Another difference was I am using DC filaments on the first gain stage in each preamp so that is what the small board is for. Also I am using cathode biasing and fixed biasing and have switching in there with relays and some solid state components to do channel switching too, so it is pretty involved. With the DC filaments and the ground buss on the board only touching the chassis at one point, and everything soldered to it in order like a galactic ground system, this amp is the quietest one I own, and I've made quite a few projects and have quite a few commercial products like an AC30, Sunn 2000S, Silvertone 1484, etc.


Whenever I get back to this amp it will look quite a bit different on the inside but you can get an idea of what I was talking about as far as wire runs and not running wires from the board to the sockets all jumbled right next to each other. Aside from the PI section in this amp, the grid runs from the board to the sockets are maybe a couple inches, whereas in your amp they are like 6 inches or more...hard to tell from the pic though. In your amp you may be able to get everything working well with that layout, but down the road you can do some thinking about how you want to lay stuff out and use some of the Hoffman examples or refer to Fenders and come up with something that will reduce the chance of having problems.


Greg

[SoundmasterG]

I am using the 12AU7's in the V1 and V4 PI stages, yes.

I just did some more troubleshooting, and I'm suspicious the issue is with the OT.  I had the power tubes out, and can't get the weird issue to happen at all.  I put in Two tubes, and ti works fine as well on the scope,

As soon as I put in all 4 tubes it starts freaking out.  Additionally I tried swapping tubes, not only in and out only two at a time, but also the socket the tubes were in (keeping them to their own half of course) so that if I had left side tube 1 and 2, first time I did 1,2 then 2,1 and with just 0 1, 1,0, 0,2, 2,0 for each case (i.e. 0 meant socket empty) and mirrored the same combo on the right side as well, I never once got the issue with just two tubes in, but eveyr way I try with 4 it does that stupid noise issue. 

I am also suspicious per something that sluckey said about the OT, so I tested the impedance of each half, and I get 19.8 ohms on one side and 16.7 on the other.   I checked and found someone selling them and lo and behold I see this:

http://www.etronic-parts.com/product_info.php/info/p1090_VOX--Output-Transformer-AC100CPH.html

they're supposed to be 18 ohms a side, is a variance of 2 ohms enough to matter?  Is the OT what's wrong here?

~Phil

You're just checking resistance with your meter? It is normal for the OT to have a mis-balance like that between sides....almost every one I have ever measured is off a bit. 18 ohms a side is what they aim for but barely any difference in the actual amount of wire on each side will make for a difference in your readings...it doesn't take much wire to give 2 ohms difference.


Hmm, so swapping around with the power tubes has no issue with two no matter the pair? I was figuring you would have a problem with one of the tubes in the quad. I am still suspicious of those tubes. Do you have a tube tester? Sometimes it will help to bring them up on a tester and see if one is way out from the others or if it has a short or something. Do you have any other EL34's or KT77's? You could try 6L6GC's also if the socket is wired appropriately if you had any of those laying around.....but really, try it without the power tubes and go through each stage with the scope and signal and make sure you can verify that there are no weird distortions or wave forms on the scope and you have clear amplification from stage to stage. Don't worry about playing it...and in fact if you have a resistor load, do your troubleshooting while connected to that instead of a speaker. It will give more predictable behavior.

Greg

[pompeiisneaks]

These tubes are brand new and the second full set I've had in it (The first set seemed like soemthing was off so I swapped two out, got two more from tube depot.com and they had pins 2/3 shorted on one side, so I burnt out the entire set of tubes heaters... so I just put in a new set this week).

Since I can literally try any two of the tubes at a time and not get the issue and ONLY when all 4 are in, it makes me think it is not the tube.  The only other thing I'm slightly suspicious of is that one of the sockets seems a tad loose to put the tube in, but I don't think that would cause this, it would be intermittent and fail at random causing some odd behavior.   As for the new power tubes, they were apex matched all 4.  I don't think it is the tubes.

Maybe I should try and tighten up the pins in the socket?

I'm just at a loss.  It definitely seems to confirm that the issue is either in the power stage or the OT.  I can't find anything specific to a socket or a tube so far, so I suspect the OT.

[shooter]

I was getting about 500V input

I'd verify this, 50v would be more *inline*

I never once got the issue with just two tubes

So ALL tubes in ANY 2 sockets, works?

[pompeiisneaks]

Correct, I kept trying every variation I could of tube and socket, but only in pairs, I.e. Tubes 1 and 2 on the left side always, and tubes 3 and 4 right side always.  I didn't specifically see any reason to swap tube pairs to the other side. 

Effectively, I did, on a half (and repeated this on the other half) where I tried tube one in socket one and two, and tube two in sockets one and two, then repeated this same for tube swap on the right side.  Nothing.  I pot in four and bam, it re appears

Phil

[SoundmasterG]

Definitely make sure the sockets are adjusted correctly on all pins. If the pin 5 is not making contact, then that tube would want to run away since it wouldn't receive its negative bias voltage, or receive it intermittently. That could very well be the problem.

Greg

[pompeiisneaks]

I had carefully gone through and re-soldered all the connections at a previous point in this thread.  I also, as a sanity check, did a continuity check from the wires before the turret right before the pins, to the base of the pin itself and get 0 ohm resistance.  The connections are all good there.

Phil

[SoundmasterG]

I'm not talking about the connections on the solder terminal side of the sockets. I am talking about the part that contacts the tube pins when you insert the tube into the socket. Make sure all the sockets are adjusted correctly so all the tube pins contact the contacts inside the sockets. When you insert the tubes into the sockets, there should be some resistance and all pins should contact the socket. If any pins are not adequately touching the socket, but especially pin 5 where you negative bias voltage comes in, then you will have LOTS of problems getting the amp to run stable as it will not get its bias and will run away. You can use a dental tool or very small screwdriver to adjust the socket terminals so they will grip the tube well when you insert the tube. Ceramic sockets are harder to adjust than micalex or other plastic sockets, but it can be done.

Greg

[pompeiisneaks]

On that may be an issue with the first tube.  It seems really loose and easy to pull the tube out.  I'll give that a try.

Phil

[pompeiisneaks]

That may have eliminated the really bad noise, but the mV readings on the cathode are still all over the place. 

Maybe I'm doing it wrong? 

Do you check the bias there with nothing going through the amp?  Or under sine wave load?  Maybe that is why I'm getting of readings?

Phil

[pompeiisneaks]

More of that weird one sidedness.  Back to just two tubes in the amp.  No popping like I've had (I pulled the other two because it was so bad almost every second now).  It's gone, but the bigger problem is that I can now measure the voltage on pin 8 and I get about 50mV on one (It has a 1.5 ohm resistor so that's about the ballpark for 70% .050/1.5 = .033mA at 480V = 16 Watts)  the other tube on the right half is at 13mV! (it is at 1.1 ohms, so .013 / 1.1 = .012 mA and X480 = 5.67 watts. )

That seems like a pretty extreme imbalance on those two tubes no? 

What would cause that imbalance? 

The imbalance follows the socket, not the tube, I switched the two of them places and I get the identical same values on the same sockets. 

Edit:  Side note I tightened the pins on the first socket and it now is quite hard to get in and out like the others.
~Phil

[sluckey]

Swap positions with the two tubes. Does the low reading follow the same tube?

[pompeiisneaks]

The imbalance follows the socket, not the tube, I switched the two of them places and I get the identical same values on the same sockets. 

From my last post...

it stays with the socket, does not follow the tube. I realize the wording isn't right... I should have said the imbalance "STAYS" with the SOCKET :).. I did swap them and it the tubes don't matter.

[sluckey]

OK, so now swap the OT plate leads. Does the low current reading move to the other socket?

[pompeiisneaks]

So I just desoldered the two plate leads from the OT.  (I left the jumper to the second socket as it was, because the pairs should still be left two right two) and re-soldered them to the opposite side of the tube pairs.  I started it back up, and yet again, the sockets contained the imbalance, not the tubes and now that I've done that NOT the OT.  Its something coming from the PI right?  I'll go over the circuit and see if I can figure out why that would be imbalanced.

~Phil

[Willabe]

To be clear, the low current reading moved to the other side/other 2 power tube sockets, when you swapped the OT primary wires? 

[pompeiisneaks]

No, the low current reading stayed with the socket.  Meaning its always been low on the right half of the tube pair. Swapping tubes didn't change the location of the low reading.  Swapping the sides of the OT didn't change anything (significant.  Before it was 50mV left tube , 12 mV right side tube , after swapping the OT to opposite halves of hte tubes, I did it again and get 50mV left tube  15mV right tube). 

I hope that makes sense.  If I did it with say just two tubes, two sockets and two sides of the OT, then let me call it LT/RT for tubes, LO,RO for output transformer and then LS/RS for sockets.  Here's what I got:

LT/LO/LS 
50mv
RT/RO/RS
11mV

Then
RT/LO/LS
50mV
LT/RO/RS
11mV

Finally
RT/RO/LS
50mV
LT/LO/RS
15mV

so it stays with the side/socket, not the transformer or the tube

I just went over the layout, and the readings on each one and I don't see anything at all that seems off/wrong. 

Looking at the schematic, (reattaching here, with a close up of the layout around the PI).

Does anyone see anything wrong there?  I don't.  I've gone over it again and again.  I also can't see any components that don't read right.  The R19, R20, R21, R22 all read correct.  C11 and C12 have consistent decent looking ESR ratings on my ESR meter.  I also checked C13 and R15 and see nothing wrong there....

~Phil

[pompeiisneaks]

Here's an older photo of the PI section.  It's in identical shape still, but I've resoldered all of it at one point to get the connections better, etc.
 
The yellow wire on the left middle, near the green wire goes off to the grid of the pin 2 of the first half of PI.  Then the two red wires just directly above it next to it and one over are the output from pins 1 and 6 of the anode on the tubes.  That goes off to B+.  the same lines, though also have dual wires coming from pins 1/6 to the left and right halves of the PI mirrored section.  Then the blue lines leave to go up to the power tubes.  I've also measured the voltages at those blue wires and see a difference there as well. 

I have the input to the grid going off the yellow wire at the top right most section where the resistor and capacitor connect in via a jumper wire to the lower section of the blue wire output through those two 1M resistors (R19/R20).  Nothing seems wrong, but something must not be right.  (BTW there is a jumper underneath that red wire on the bottom between R19/R20 , you just can't see it because it is covered).



~Phil

[Willabe]

In your drawing pin 8 is not connected.

[pompeiisneaks]

Oh yeah that's just an error on the layout drawing, I've jumpered at the socket pins 3 and 8. 

I'll fix that now.

~Phil

[pompeiisneaks]

Here's the best picture I have of the PI, you can see the black wire jumpering the two.

~Phil

[SoundmasterG]

That may have eliminated the really bad noise, but the mV readings on the cathode are still all over the place. 

Maybe I'm doing it wrong? 

Do you check the bias there with nothing going through the amp?  Or under sine wave load?  Maybe that is why I'm getting of readings?

Phil

You should know that on the particular socket that you adjusted, any tube that was in there while you were doing your testing likely may not have been receiving its negative bias voltage on pin 5 that would have kept the tube in check, so that tube may have been cooking itself the whole time. Something to be aware of.


When you check the bias, all controls should be at zero and there should be no signal going into the amp. You should also have it on a resistive load instead of a speaker as a speaker causes differences at different frequencies and the resistor load allows all frequencies to perform the same. If you try to check the bias when you have a signal going into the amp, the cathode reading that you get across your 1 ohm resistors will change. Setting the bias is done when the amp is in an idle condition, which means no input to the amp. Try it again with no input signal and all controls at zero and see if the readings are the same.


Greg

[Willabe]

In the pic you just posted, the board mounting screw looks very close to a turret, might just be the pic angle?

[pompeiisneaks]

I went and looked and it is a bit close, so I removed the screw (5 screws is likely fine) but it didnt change anything, still imbalanced.  Also when I check the voltage of the lines coming out from the PI one side is -40V the other is -37 V so theres a 3 V difference from that as well, (not sure if that makes any difference)

~Phil

[sluckey]

If you are concerned that the pi is somehow causing the problems with your output tubes, just disconnect the coupling caps from the pi plates. Then you will be dealing only with the output tubes, OT, B+ supply, and bias circuit.

[pompeiisneaks]

By disconnect you mean remove and jumper over right?  Because they're what connects into the inputs of the Power amp right?  Does that make me lose the dc rejection though?  Or should it be pretty nominal at this point of the circuit? 

~Phil