AB763 Stunflower

[romberg]

So, I think I have my modified normal channel sorted out. I auditioned a handful of bypass caps and ended up settling on a 0,68u on the second stage. This value seems to keep the bass night and tight. Next I added a 100k pot across the 10k cathode resistor and played it a while biased from cold to warm. I found that the cold bias (10k at around -4v) was just too much. It provided so much clipping/distortion that you could not really clean it up much with the volume pot on a guitar. Plus it did not sound real pleasing. I ended up liking a bias of -1.5v which a 1.2k cathode resistor provided. I'd now say that this normal channel is at least as mean as a cascaded marshall 2204. Mission accomplished .

But I've found another bug to work out. There is this extra noise which I had mistaken fir the second stage of my input channels starting to breakup. But It is not. I think it is being caused by the reverb driver tube. The signal looks clean on either channel on the other side of the mixing resistors. But at their junction I see:

rdrive.png

Since the signal is clean before this the only other thing that makes sense to me that is connected their is the grid on the reverb drive tube. I've noticed a few interesting things:

• The signal is not distorted if either preamp channel is set low. Then it starts to distort when either channel is brought up to a certain voltage.
• Moving wires associated with the drive tube with a chopstick made no noticable change in the distorted waveform.
• Physically moving the reverb tank made to difference.
• Unplugging the send to the reverb tank totally and completely cleans up the signal!

I think that the reverb driver tube is being overdriven and this is causing some sorta signal to be injected onto the grid (and thus into the channel mixer) via grid current or something. I did not bust out my high voltage probe to confirm this but the input signal you can see above looks more than enough to be over driving this stage. My goal is to make this noise go away. I guess I don't care if it enters the reverb tank but I don't want it in the normal signal path of the preamp.

I'm thinking of using Merlin's universal cureall and adding a grid stopper to the reverb driver. Does this sound like I'm on the right path? This is the first reverb circuit I've done and I'm just kinda guessing at things.

Thanks,

Mike

[romberg]

So, much for the grid stopper idea. I tried various values from 39k to 100k and they made no difference at all. I'm starting to think that I may have stumbled onto the "blackface sound". This distortion is there on all channels from about 5 on the dial on up. And it is there all the time. Even when the reverb is off via the foot switch or pot it won't matter. The signal gets this extra distortion from the reverb driver.

I'm now thinking voltage dividers before the reverb driver tube. The signal is clean up to around 8Vpp but can rise to 32Vpp when the tremolo channel is cranked. So, maybe add a 4:1 voltage divider before the reverb send. But maybe this will kill the whole blackface vibe? The cleans may get even cleaner with more headroom. But then the reverb character may change since it is not being driven as hard?

Or maybe something is fubar with my reverb transformer? Swapping the tube with another 12at7 made no difference.

Mike

[romberg]

So, I installed a 270k/820k voltage divider in place of the 1m grid leak on the reverb driver tube. And it did totally and completely clean up the messy signal that was being injected into the audio channels. However this (unsurprisingly) did have an effect on the reverb itself.

The reverb knob was now much more useable over it's entire range. I had already used a log/audio pot for the reverb but with the stock 1m grid leak even it came on pretty fast and furious. With my voltage divider knocking off 75% of the signal the whole range on the pot was now useable. I only use enough reverb that you miss it when you turn it off so this extra range on the pot was a good thing. However, driving the tank less hard seems to have killed the whole *vibe* of the reverb. It just did not sound as good.

So, clearly a compromise was needed. After searching the net, I found that replacing the 1m grid leak with a pot (the "dwell" control) was a thing that is old hat to many fender tweakers. So, I wired in one temprarily to experiment with differing ratios on the divider. I found something interesting.

I did not have to reduce the signal too much to delay the onset of reverb driver distortion until after the PI was distorting. And once the PI was grinding away with overdrive you don't see much of the distorted preamp signal ontop of it. I settled on only kncking down the signal into the reverb driver by 20%. Below is a scope shot of the trem channel on 10 with a 100mVrms signal with this setup.

dwell.png

Here the yellow channel is on the grid of the third gain stage. Note that on 10 neither of the first two stages of the trem channel distort at all. This is all due to the driver tube verdriving. The amp output here is in blue and just slightly shows the distortion from the reverb driver. The master is turned down so this is all PI overdrive.

The signal into the tanik is still strong enough to keep the reverb sounding like the stock setup. So, I think I'll wire in a permanent 220k/820k divider and go with that (no room for an extra dwell knob). And then the extra reverb overdrive will still be there but only after the power tubes and or PI have started to overdrive. So, you can't hear it.

This must be part of the reason the silverfaced reverb amps with a master volume had such a bad reputation. If you put a pre PI master on there then you are gonna see that yellow signal at the amps output and the PI wont clip till much later. It seems that the CBS engineers tried to fix this by killing the gain in the reverb driver tube. Which I'm guessing had a negative effect on the reverb itself.

Mike

[romberg]

I've just ordered a bag of vtl5c1s and am getting ready to convert my stunflower over to a Trem-O-Nator. When it is possible, I like to do a bit at a time and test as I go. So, I'm thinking of doing the conversion in two steps. The first would be to replace the driver stage and bug with the vtl5c1 (keeping the stock oscillator stage). Something like this:

trem-o-nator-a.pdf

Then, phase two would be to upgrade/change the oscillator to trem-o-nator specs.

Few things I'm unsure about with trying the vtl5c1 with the stock oscillator.

• Should the cathode resistor be changed to 10k? I think the triode will be biased mostly/completley by the LED's forward voltage. But this is shady territory for me so I could easily not be understanding something important.
• Should the 220k plate resistor be bumped up to 470k to keep the current down and avoid frying the vtl5c1?
• Not sure about keeping the whole stock footswitch and raw bias voltage setup to start/stop the oscillator. I'm kinda planning on keeping this even with a converted
  trem-o-nator oscillator as I don't have a switchable rca jack for the pedal (do they even make such a thing?).

Any advice on all this would be greatly appreciated. Worst case, I'm gonna have a bag full of 'em and can let the smoke out of a few while learning .

Mike

[Ten Over]

[*] Should the cathode resistor be changed to 10k? I think the triode will be biased mostly/completley by the LED's forward voltage. But this is shady territory for me so I could easily not be understanding something important.
[*] Should the 220k plate resistor be bumped up to 470k to keep the current down and avoid frying the vtl5c1?
[*] Not sure about keeping the whole stock footswitch and raw bias voltage setup to start/stop the oscillator. I'm kinda planning on keeping this even with a converted
trem-o-nator oscillator as I don't have a switchable rca jack for the pedal (do they even make such a thing?).
[/list]

Change the cathode resistor to 10k or larger, like 47k. The 2k7 cathode resistor will steal current from the LED.

There is a large amount of variation amongst VTL5C1's. Some work fine with a 470k plate resistor and some need the extra current that a 220k plate resistor provides. I seriously doubt that you can fry one of those VTL5C1's.

The stock footswitch setup is fine. It puts the triode firmly in cut-off so that the LDR has a very high resistance.

[Ten Over]

I've just ordered a bag of vtl5c1s and am getting ready to convert my stunflower over to a Trem-O-Nator.

Where did you order those from (please don't say Amazon).

[Ten Over]

You're not going to want that SPST switch above the 50kRA pot. It will make the Vibrato channel louder when open. If you want a switch, you can put it across the foot switch jack or in the path of the LED.

[sluckey]

Few things I'm unsure about with trying the vtl5c1 with the stock oscillator.

• Should the cathode resistor be changed to 10k? I think the triode will be biased mostly/completley by the LED's forward voltage. But this is shady territory for me so I could easily not be understanding something important. The LED is all you need on the cathode. The purpose of the 10K is to keep the oscillator alive at a reduced current when the LED is switched off. This insures an instant startup when you stomp the footswitch.
• Should the 220k plate resistor be bumped up to 470k to keep the current down and avoid frying the vtl5c1? Hmm. I never even thought about that since I chose to use the 6G16 Vibroverb oscillator. May as well use the 470K. The oscillator will work with either value.
• Not sure about keeping the whole stock footswitch and raw bias voltage setup to start/stop the oscillator. I'm kinda planning on keeping this even with a converted trem-o-nator oscillator as I don't have a switchable rca jack for the pedal (do they even make such a thing?). You don't need that stuff to kickstart the oscillator because that 10K keeps the oscillator running all the time. And you can use your same rca jack. But you will need to keep the footswitch plugged in to enable the trem-o-nator. This is the same as your current AB763 tremolo. I used a switchcraft 12A switched jack only because I was modifying a Bandmaster which uses a single button footswitch and I wanted to be able to have trem without a footswitch. Plus my single button footswitch uses a 1/4" phone plug .

Any advice on all this would be greatly appreciated. Worst case, I'm gonna have a bag full of 'em and can let the smoke out of a few while learning . I highly recommend not to do the mod in stages like you plan. It's a fairly simple and straightforward mod that's easy to undo if you get hung up. I hope you found a stash of Vactec/Vactrol optos. A few people have had issues with the Xvive optos.

I'll be following along with interest. This should be an easy mod and since you have an AB763 circuit the 50K-RA INT pot will interface exactly as your current INT pot.

[romberg]

Change the cathode resistor to 10k or larger, like 47k. The 2k7 cathode resistor will steal current from the LED.

There is a large amount of variation amongst VTL5C1's. Some work fine with a 470k plate resistor and some need the extra current that a 220k plate resistor provides. I seriously doubt that you can fry one of those VTL5C1's.

The stock footswitch setup is fine. It puts the triode firmly in cut-off so that the LDR has a very high resistance.

Thanks for the advice! I was wondering about the 10k cathode resistor. I think it may only need to be there if one uses the trem-o-nator footswitch setup which disconnects the ground to the LED. If the stock fender switching is used then maybe the LED alone could be used on the cathode? I'm gonna play it safe and put in a 47k to start with.

Mike

[romberg]

I've just ordered a bag of vtl5c1s and am getting ready to convert my stunflower over to a Trem-O-Nator.

Where did you order those from (please don't say Amazon).

Umm.... Well... I won't say it then but that is where they were ordere from. They did say they were vactrols. BUt I'd not be surprised if they were something else. I'm guessing I should expect the worst and order at least one from another source? I think I saw that Antique Electronics had some and I need a few tubes anyway. So, I'll order a backup.

Mike

[romberg]

You're not going to want that SPST switch above the 50kRA pot. It will make the Vibrato channel louder when open. If you want a switch, you can put it across the foot switch jack or in the path of the LED.

I've already got the switch on the amp. And it does do exactly what you say. It reduces the loading on the preceding stages and bumps up the gain a bit. I kinda like it for now. Once I add the freed up triode as another gain stage, I might revisit this switch.

Mike

[romberg]

I highly recommend not to do the mod in stages like you plan. It's a fairly simple and straightforward mod that's easy to undo if you get hung up. I hope you found a stash of Vactec/Vactrol optos. A few people have had issues with the Xvive optos.

I'll be following along with interest. This should be an easy mod and since you have an AB763 circuit the 50K-RA INT pot will interface exactly as your current INT pot.

I'll take this advise seriously. I'm mostly worried about the wires I ran under the board and the foot switch setup. But since I may be able to use the stock fender foot switch along with the trem-o-nator oscillator, I very well may just dive in and do it all at once. I'm gonna draw it all up while I'm waiting for parts to arrive...

Mike

P.S. Thanks again for sharing your design for this!

[Ten Over]

I've just ordered a bag of vtl5c1s and am getting ready to convert my stunflower over to a Trem-O-Nator.

Where did you order those from (please don't say Amazon).

Umm.... Well... I won't say it then but that is where they were ordere from. They did say they were vactrols. BUt I'd not be surprised if they were something else. I'm guessing I should expect the worst and order at least one from another source? I think I saw that Antique Electronics had some and I need a few tubes anyway. So, I'll order a backup.

Mike

A guy on another forum had trouble with some of those that he had ordered on Amazon. I ordered in some, waited a few days and ordered some more in order to possibly have two batches. After testing them, none of them were anywhere near a VTL5C1 and were useless for the trem-o-lator. Maybe you will have better luck. Ordering a couple from AES is a good idea.

[Ten Over]

If the stock fender switching is used then maybe the LED alone could be used on the cathode?

Yes.

[romberg]

A guy on another forum had trouble with some of those that he had ordered on Amazon. I ordered in some, waited a few days and ordered some more in order to possibly have two batches. After testing them, none of them were anywhere near a VTL5C1 and were useless for the trem-o-lator. Maybe you will have better luck. Ordering a couple from AES is a good idea.

It turns out that the AES parts are XVive. In fact I can't seem to locate anyone (that seems reputable) selling a vactec vtl5c1. In what way are the newer parts deficient for use in a trem-o-nator? I will test the handful I get before tearing down the stock tremolo.

Mike