Switchable SS/Tube rectifier question

[osing]

Howdy,

I would like to add a switchable SS rectifier on my Hoffman AB763 Super Reverb. I found the attached diagram posted by Sluckey in an old thread started by bnwitt when he was working on an SRV Vibroverb project (http://el34world.com/Forum/index.php?topic=3238.msg27872#msg27872). I like the idea of making the switchable option part of the standby switch. Having read some other posts as well, I gather for a 2x6L6GC amp I should use two pairs of 1N4007 diodes in parallel rather than just two in parallel to get a higher overall voltage rating?

When I look at the Fender Custom Vibroverb schematic found in the thread referenced above  http://www.prowessamplifiers.com/schematics/images/64%20Vibroverb%20Custom%20Schematic.pdf I get confused as I see two diodes between the PT and the rectifier tube (lower right hand corner of the schematic) as well as four diodes near the power tubes (upper right hand corner of the schematic). I've also seen suggestions of adding capacitors to help even out the voltage on the diodes.

I guess my question is - is there any real reason not to go with the simple approach in the attached diagram?

[shooter]

Mesa boogie uses a dual rect in their solo head unit, you can find the schematic here in dougs schematic achieve. 

[sluckey]

I should use two pairs of 1N4007 diodes in parallel rather than just two in parallel to get a higher overall voltage rating?

The diodes would be in series, not parallel.

The '64 Custom Vibroverb circuit is really simple too. The attached diagram shows both circuits for easy comparison. I prefer the single switch. The red note is a good idea and very easy to do.

Those diodes in the output tube section have nothing to do with the selectable rectifier circuit.

[lego4040]

Very cool and a great way for some Umphhh but wouldn't you have to Bias after wards? Cathode bias will do it itself and I guess you can just flip the switch and play but if you had a fixed bias you'd have to do yourself wouldnt you?

[osing]

Thanks guys,


I think what Sluckey is saying is that when using the tube rectifier in the '64 Custom Vibroverb (see schematic link above), there is an added resistor R79 (2.4k) between the bias pot and the "bias diode" (there's probably a better name for it than that). When the SS rectifier is switched on, the DPDT switch not only bypasses the rectifier tube but also this 2.4k resistor, thereby changing the bias voltage.


If I am understanding this correctly, you can play around with this resistor value (or perhaps calculate it if you're capable - I'm not) so that you have a proper bias regardless of which rectifier you're using.


Sound right?

[sluckey]

If I am understanding this correctly, you can play around with this resistor value .... so that you have a proper bias regardless of which rectifier you're using.

Exactly. You could even use a 5K or 10K pot and have a two bias adjustments, one for tube, one for SS.

However, if you don't need the amp tuned to NASCAR specs, forget about switching the bias, and follow the KISS rule. Just set the bias for your optimum flavor in SS mode and accept that the amp will be biased a little bit cooler when in tube mode. There may not be enough difference switching between SS and 5AR4 to matter.

[HotBluePlates]

Also, the setup on the Vibroverb Custom has the solid-state diodes do the actual rectifying all the time, and allows you to place the tube rectifier in series (for added sag) or not (for SS rectifier sound).

If you look closely, the tube rectifier plates are in parallel at all times. So your GZ34 will have half the impedance it normally would exhibit, and less sag. If it were me, I'd probably stick a 5U4 in there to get some of the sag back.

[osing]

Thanks again guys. I might start out trying Sluckey's version so that I can eliminate the diodes completely with the flick of a switch, and I might even try adding the second bias pot as well in place of a fixed resistor (i.e. first adjust the main bias pot in SS mode, then switch to tube mode, and adjust the second bias pot which is in series with the first). Worst case I undo it.


I am now curious though how the Vibroverb version would sound, maybe I will get around to trying both. HBP - I was confused about how the Vibroverb was wired because I could not figure out how the diodes would get switched off, so your explanation made it a lot clearer.

[osing]

Hi again,


I got the parts I need from Doug for adding the switchable SS/Tube rectifier mod. As I was thinking more about implementing this mod via the ON-OFF-ON DPDT switch option as mentioned by Sluckey above, I tried to do some research on how the tube rectifier might be affected while in SS mode. My Super Reverb before I do this mod has two 70uF caps between the rectifier tube and the existing standby switch similar to the original AB763 circuits. I've seen various opinions about whether to place these "reservoir caps" before or after the standby switch. Some suggest it is beneficial to the rectifier tube to have these filter caps before the standby switch because as the rectifier tube heats up these caps are charged more slowly and put less strain on the rectifier tube.


With the SS/tube rectifier standby switch as per Sluckey's diagram above, I need to put all the filter caps after the standby switch. Whether it matters or not, when I want to use the tube rectifier, I guess I can place the standby switch in the "tube" position before turning on the main switch. This way, all tubes are heating at the same time and there won't be any sudden rushes of current. When I choose to use the SS rectifier from a cold start, I would always leave the standby in the center OFF position until the main switch has been turned on for a while to allow the tubes to heat up first.


I read on Merlin Blencowe's web site http://www.valvewizard.co.uk/standby.html that if a rectifier tube sits in standby for hours on end, it can suffer from something called cathode poisoning. However, it's not clear exactly how much time would be needed for there to be damage to the rectifier tube. If my amp is played for 4-5 hours in the SS position, there will be voltage at the rectifier tube but no current. Does anyone have thoughts and/or experience whether this a real problem, or more of a theoretical one for the duration in question?

Also, given the 400+ volts coming of the rectifiers, is there an issue using a switch rated for 15A 250VAC? Isn't a 250VAC standby switch common and if so, why is there no issue?

Thanks!

[osing]

I created a schematic (see first file below "Switchable rectifier option 1") for how I would end up implementing Sluckey's suggestion above with the DPDT ON-OFF-ON switch allowing me to go from standby to either SS or tube rectification and also add in a resistor (or second bias pot) to the bias circuit when in tube rectifier mode (similar to '64 Custom Vibroverb).


I also created a second schematic "Switchable rectifier option 2". I am curious if this version would sound the same. When in SS mode, the tube rectifier is still in the circuit, but the theory would be that it is not really doing much since the bulk of the current would be coming via the diodes. Is this a valid theory? Further, since the tube rectifier is still in circuit, it might possibly not get subjected to cathode poisoning (see previous post - not sure if this is even a real concern)? Finally, this mode does allow for the reservoir caps to be located on the "hot" side of the standby switch, which may or may not be beneficial to the rectifier tube.


With option 2, I could not figure out a way to do the switching without a 3PDT switch.


Any thoughts as to whether there is any value going with option 2?

[terminalgs]

Option #2 has a problem: you've got a switch in series with the choke, yet after the first reservoir cap.    you want caps on both sides of the choke, all the time. otherwise, you'll have flyback voltage spikes.  That can damage the choke (google 'flyback diode'). Option#1 is better.


this Might work (if you are worried about that condition Merlin warns about...): if your 1st filter cap is 47uf, you could put a 100uf on one side of the switch and a 100uf on the other side as well....  that way the choke isn't left with an unconnected side.  When you close the switch, total capacitance is 50uf.


personally,

[kagliostro]

I hope you don't consider this post an invasion of your tread

You asked about the diodes present on the PA area of the and also I am interested to those diodes

as far I can understand the purpose of D2 and D3 that are connected backwards  is to protect discharging to ground overvoltages

D4 and D5 seems to give a voltage reference to the cathodes of the Power Tubes, acting as a capacitor (a short) for AC

but what displaces me is the presence of the Flame Proof resistors R72 and R73 because they are in parallel with the diodes

Please can someone give an explanation to this part of the circuit ?

K

[PRR]

There is no reason to have D4 D5.

[kagliostro]

There is no reason to have D4 D5

OK, assuming that, however Fender has think they may had an action on the circuit .....

To prevent sound alteration due to the fact that there are the two 1R Fp resistors connected between catodes and ground ???

are they there to act as a short for AC signals, thing that didn't happen at 100% with a resistor ???

Are the protection diodes for some reason connected that way instead than connected directly to ground ???

I'm very confused about

Franco

[PRR]

I think the designer was just confused.

D4 D5 won't conduct until the 1 Ohm resistors (and tubes!) pass 0.6 Amps. This will *never* happen normally.

[osing]

Thanks terminalgs,


So, does the original Super Reverb AB763 design (see schematic below, although please ignore most of the red circles, they are not relevant to this discussion) have this same flaw then regarding the possible flyback voltage since it appears that the standby switch is right before the choke? Did Fender do away with this approach at some point after the blackface amps?


I did some more reading on Merlin's web site, and he suggests putting a 2W resistor across the standby switch as "both inrush surge and cathode poisoning can be completely avoided". I modified option 1 and renamed it 1b - this has a 100k 2W resistor across the standby switch. I also modified option 2 (now 2b) by adding the same resistor across the standby switch, but I also moved the reservoir capacitors to between the switch and the choke.


Between 1b and 2b, is it fair to say that they now accomplish the same thing, except 1b does so in a simpler fashion (i.e. there is no need to do the 3PDT switch as per 2b any longer since the 2W resistor in 1b will prevent the cathode poisoning which is what option 2 was originally intended to do).


Thoughts?

[DummyLoad]

attached is how i'd implement the bias PS compensation switching.


--pete

[PRR]

> possible flyback voltage since it appears that the standby switch is right before the choke?

AND the output stage.

Yes the OT is inducty but from the B+ to both tubes it is *not* an inductance (is not working push-pull, the inductances cancel). So at standby turn-on the B+ is loaded by hot 6L6 conduction. For typical bias this is like a 5K resistor; IAC it is ~~80mA of DC current. That modest choke can't "kick" with that connected.

I do not think cathode poisoning is in my top-10 list of things to worry about, not with receiving tubes in guitar amps. Interface in cut-off (or no-B+) tubes is an issue for some applications, especially very marginal computer circuits. But it wasn't really "discovered" until those early computers. Meanwhile back home many TV sets turned-off the B+ and left the heaters warm for "instant on". With longer standby periods than a stage-amp should get. These TVs usually came out of standby just fine. And in the computers, some time at full B+ restored proper operation, maybe just a few seconds. It was an issue because a computer may do a LOT of work in a few seconds and if a bit "sticks" after being off a while the results come out wrong. You will usually have full B+ for many seconds before you need to play. And if the tube is a little bit down on bias the first minute you probably won't notice.

These Fender circuits worked well. IMHO this lily does not need over-gilding.

[kagliostro]

D4 D5 won't conduct until the 1 Ohm resistors (and tubes!) pass 0.6 Amps. This will *never* happen normally.

are the two diodes there for some mixed bias purpose  (fixed + catode bias) ??


K

[osing]

So I wired up the SS rectifier as in the attached diagram. At first test, everything seems to work ok. However, I seem to be experiencing a strange phenomenon in one particular instance. I have a boost switch installed where I break the connection after the 47k resistor in the trem circuit (see attachment with red line indicating where the disconnect occurs - I had seen this in one of Sluckey's old posts). I also have a 6 lug reverb/trem jack wired so that the reverb is on but the trem is off when no pedal is connected (see attached diagram from Sluckey - the trem is grounded when there is no pedal connected and is thus off).


Anyway, with no pedal connected (i.e. thus trem is off), and the boost switch on, if I go from either SS or tube rectification to the other (via standby since the center position of my SS/tube switch is standby), I get this horrendous machine gun sound (do not confuse with the awesome sound of Hendrix' great song) which will only go away if I toggle the boost switch off to connect the 47k resistor again. I can then immediately flip the boost back on and everything is fine. However, if I again change rectifiers, the same problem occurs.


Now, if I have the pedal connected, this does not happen. I can leave the boost on and switch rectifiers with no issue, regardless of whether the trem is on or off.


I am curious if anyone has any thoughts as to why this might be happening?

[sluckey]

Don't know if this is applicable to your problem, but there's an error on that Hoffman pic you posted. The bottom side of that 47K resistor (and the other 3 resistors plus the cap) needs to be connected to ground. Are they in your amp?

[osing]

Thanks Sluckey - I did not notice that error in Doug's layout, but I just confirmed that my amp does indeed have those grounded. Except, when I engage the boost switch, then the 47K is lifted from ground. Before I installed the DPDT (ON-OFF-ON) switch and selectable rectifier, I never had any issues with this boost, and it only happens when I go from one rectifier to the other with the boost on.

[sluckey]

That's odd. Sometimes **it just happens. Maybe try a 1M in place of the 47K and use the switch to put the 47K in parallel with the 1M, rather than opening the ground path to the 47K.  Switch open would be boost, closed would be normal. Just throwing this out there.

[EL34]

Thanks for the heads up guys


I fixed the one channel and two channel drawings on the library page


make sure you hit refresh so your browser will load the new PDF file

[osing]

Sluckey,


I tried the 1M resistor in parallel with the 47k resistor, but unfortunately it did not help with this issue. I also disconnected the resistor I placed across the DPDT standby switch but that did not help either (I figured it was worth a try).


I disconnected the speakers and used a dummy load instead. Interestingly, I can hear the stuttering sound coming out of the circuit, albeit obviously not as load as with the speakers connected. I tried isolating the sound to specific components but to no avail - I even tried using a funnel to my ear 


I took some snapshots with my oscilloscope having connected the probe and ground to the tip and ring of the speaker jack at the dummy load. I did not insert a signal into the input of the amp - the snapshots show the stuttering sound only. It seems as though there are some regular large spikes whose min and max voltages do not change with the volume of the amp, but the smaller waves in between do get larger with volume. I am not sure what the frequency measured by the scope represents, but by looking at the distances between the large spikes, it seems like they happen every 55ms or ~18Hz.


I did make some observations:
1) With the intensity knob at 10 and the speed knob at 9 or below, I can get the sound to disappear. With the intensity at 10, the 0.1uF cap near V5 and the 220k resistors are shorted - see highlighted wires in the attached picture.


2) The level of the sound changes perhaps a little as I turn the speed and intensity knobs, but the frequency does not seem to change by much if at all.


3) I touched the oscilloscope probe to pin 7 of V5 and the sound disappeared. Once I removed the probe, the sound came back.


4) As mentioned earlier, the sound can be eliminated by switching the boost off and then back on again. It's like a reset.


I am not sure if any of this helps, but I would be happy to take further measurements or try other ideas if anyone has suggestions!

[HotBluePlates]

Motorboating?

[osing]

I had to look up what motorboating is, and if I understand it correctly, it is essentially some sort of low frequency oscillation that occurs due to some unwanted feedback from one stage into an earlier stage. That seems like it could be plausible. How does one determine what exactly is being fed back and why is it happening only when the boost is engaged and switching between one rectifier to another (or technically, it might be when switching from one rectifier to standby on the way to the other rectifier).


Any thoughts on what I should check?

[sluckey]

or technically, it might be when switching from one rectifier to standby on the way to the other rectifier

Just switch to standby and then go right back to the same rectifier. Do this several times. Does it still do it?

[osing]

Just tried tube rectifier mode, moved to standby, then back to tube rectifier, then standby, etc. The problem still persists. There is an exception, though. If the Speed knob is set to 8 or below, there is no annoying sound. However, all I need to do is turn the knob up past 8 and the sound reappears. If the Intensity knob is at 10, then the sound will only be audible with the speed knob higher than 8. If I turn the Intensity knob down just a little bit from max, then the sound will remain as I turn the Speed knob down lower than 8 as well.


I could swear that when I was working on this earlier that I could turn the boost off and back on and then eliminate the sound until I changed rectifiers. Now it seems like this is not possible - i.e. the sound will disappear when I turn off the boost, but it will come back as soon as I turn the boost back on even without switching rectifiers. Either I am going crazy or the problem is not consistent.


I can confirm, though, that the problem does not occur if the pedal is connected. The sound never happens if the pedal is already connected prior to startup. Similarly, if I plug the pedal in after startup while the problem is audible, the sound disappears as soon as I start plugging it in. As I reminder, I have the tremolo wired to be off with no pedal connected.

[osing]

I had a chance to look at this a little more today....


1) With the oscilloscope I got the attached waveforms at the first filter cap with and without the annoying stuttering sound. I assume the low voltage 119Hz signal is normal at the first cap, but the second one where the 119Hz signal appears to be modulated or riding a lower frequency wave at ~20Hz probably is not normal?


2) With my meter set to AC, and with the stuttering sound present, I measured AC signals in the 40mV - 200mV range at the plates of V1, V2, V4, and V5a. I also measured AC signals on the filter side of the plate resistors on the order of ~40mV (in all cases lower than at the plates). With no stutter sound present, I measured only 2-4mV at either side of the plate resistors. I would have expected 0V but perhaps this is just residual AC ripple.


3) Since I am able to eliminate the sound in some cases by decreasing the Speed setting I am effectively increasing the resistance to ground to something higher than 100k at the highlighted point where the red arrow points to in the attached layout. Similarly, I am able to eliminate the sound when touching a meter or oscilloscope to Pin 7 of V5 which is the point indicated by the blue arrow. I don't really understand the trem circuit, and I don't know if this gives anyone some further clues or not.


4) I also do not understand why connecting the trem pedal makes a difference (I cannot reproduce the problem with the pedal connected). When not connected, the point indicated by the green arrow is shorted to ground via the reverb jack (as a reminder, I have the trem wired to off when there is no pedal). When the pedal is connected, this point is shorted to ground when off and opened when on.


I don't know if any of the above bullets gives any further clues, but if you have any thoughts, please share!

[sluckey]

Were you having any problems with this amp before you installed the dual recto circuit? If not, then completely reverse the dual recto mod and see if it works again.

Pull V5 (trem tube) and lay it aside. Does the problem go away?

You may have some layout issues. Can you post some hi rez pics?

[osing]

The amp was indeed working fine and sounded great before I added the switchable SS rectifier.


I removed V5 and was not able to recreate the problem. I then plugged it back in again and disconnected the SS rectifier from the standby switch by removing the wire located at the tip of the "V" shape formed by the diodes on the rectifier board. I turned the amp back on with the tube rectifier and could not recreate the problem. It seems like the issue is somehow related to the SS rectifier, but I don't understand how that is possible when I have heard the stutter sound in both SS and tube rectifier modes. Obviously, I am missing something.


I have attached some photos - please let me know if you want more close ups of specific areas. I am sure there are some possible lead dress issues whether related to this problem or not. You will notice several switches near V1 and V2. The first one is for a normal channel disconnect. The next two are for switchable cathode resistors/caps. The fourth one is the simulated tremolo disconnect boost. I am using this amp as a way to learn and try out different things. I might be a little "switch happy", but I find that I cannot do a proper A/B comparison otherwise. My intent is to build a more compact one channel amp with the values I find most pleasing to the ear. Anyway, just wanted to point these out and also make it clear that these were installed a while ago with no apparent issues (although there now appears to be something going on between the SS rectifier and the the trem "disconnect" boost).


Thanks again!

[osing]

2nd pic

[osing]

And 3rd pic

[osing]

I started playing with the bias to see if changing it would somehow get rid of the stuttering effect. it did not. However, I made another observation. I found that when the stuttering effect was present, the bias current on my V8 6L6GC would increase by 7-8mA, whereas on V7 there was no change. As my power tubes are already 7-8mA apart to begin with, the difference with the stuttering effect was as high as 16mA.

Anyway, I swapped my V5 (JJ 12AX) tube for an old Grooves Tube 12AX that originally came with the Reissue SR, and the problem disappeared. I don't know why I didn't try this before, especially after Sluckey suggested removing V5 to see if the problem disappeared. Either way, I am happy the problem seems to be gone.

It does leave me with some questions I still cannot figure out. If the tube was bad, why was there no problem after I added the switchable SS rectifier as long as the reverb/vibrato pedal was plugged in? Why would the problem go away as soon as I touched pin 7 of V5 with a meter or oscilloscope probe? Finally, why was there no issue prior to adding the SS rectifier? Even after I added the mod, as soon as I disconnected the diodes from the circuit again, the problem went away.

Perhaps Sluckey said it best **it just happens!

[2deaf]

Kinda erodes your skepticism of Voodoo, doesn't it?

as soon as I touched pin 7 of V5 with a meter or oscilloscope probe?

Here's a self-taught lesson that may be of benefit to some of you and spare you the experience of discarding a time-consuming pile of data.  That probe can radically alter a circuit and give you erroneous results.  Once you attain whatever reading you want, disconnect the probe before taking readings elsewhere.

[sluckey]

I wonder if your TRS footswitch jack is wired correctly? Especially the grounds. I'd love to see how it behaves if you totally remove that jack and any wires that were connected to it. Would you mind doing that?

[osing]

if you totally remove that jack and any wires that were connected to it. Would you mind doing that?

Sluckey - I can definitely do that - I assume you do not need me to add the old jack, but just disconnect it, right? That should leave both the reverb and vibrato on, I believe?

[sluckey]

I assume you do not need me to add the old jack, but just disconnect it, right? That should leave both the reverb and vibrato on, I believe?

Yes and yes. I do want you to remove the wires also rather than just leave them dangling.

[osing]

Sluckey,


I plugged the JJ tube back in, made sure I could still get the stuttering effect, which I did. I then removed the switching jack and the three wires completely. I was not able to get the same effect again. Thus, there seems to be something with either how the jack is wired or something with the jack itself that is causing this. I checked the wiring again, and it seems valid to me. I attached a closeup of the switching jack in case you want to take a look. Does anything seem odd to you?