Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Skydragon on March 01, 2018, 07:22:01 am
-
I've recently built a single-channel AB763 type amp and noticed that when I turned the volume up beyond 6 or 7 that I'd get a horrible (crackly/harsh) distortion sound. This would appear on the sound even with the reverb turned to zero.
After doing some internet research I discovered that the problem was the reverb driver tube being overdriven and in turn, affecting the B+ supply to the preamp tubes. It seems some amps suffer more than others in this regard (?)
The original AB763 design calls for a 2,200 Ohm resistor and a 22uf capacitor on the cathode of the reverb driver tube.
The cure for me was remove the 22uf cap. The result was the harsh distortion disappearing. fwiw I also intend to reduce the cathode resistor down to 1,500 ohms, when i next have a spare 5 minutes on the amp.
It is interesting to note that later fender designs, use a far lower value resistor and no capacitor on the reverb driver cathode.
This got me thinking....how many people out there have Deluxe Reverbs that don't sound good at high volumes, due to the reverb driver circuit distorting (and the owner not realising that is the reason)??
-
This is the first time I've heard of this problem. What kind of tube are you using for the driver? Which power supply node is feeding the driver?
It is interesting to note that later fender designs, use a far lower value resistor and no capacitor on the reverb driver cathode.
Which later models are you referring to?
-
This is the first time I've heard of this problem.
Me too! Seems odd?
With respect, Tubenit
-
Sluckey and tubenit beat me to it.
Them saying it's the first time they've heard of it is pretty substantial.
There may be a separate issue that you have that has been masked by removing the bypass cap.
I'd have to put the cap back in and try to find the real issue.
Do you have a schematic for this build?
-
I was surprised to read this too. I always presume the old original designs work 100%. I’ll try and post some links later.
Before I read about the problem and cure, I tried a number of things including altering the B+ to the reverb. It didn’t make any difference.
If you look at the older backface designs and the evolution through silverface schematics you’ll see how fender evolved the reverb driver to avoid the reverb driver being overdriven. Likewise (for different reasons) the reverb return also has a few changes.
The amp is an exact copy of a AB763 with the normal channel and vibrato removed. I will post an exact schematic when back home in a few days.
The long and short is however is that the standard circuit will overdrive the reverb driver at high volumes
-
To add more detail
I’m using the amp in a ‘non standard’ way by having the power stage on a VVR scaling setup and a pre-PI master. Which means I can have the preamp running a high volume and hear that audio whilst at a low volume. This means it’s easy/easier to hear an issues with the preamp.
Some of the posts I was referring to from other people were using standard amps
-
power stage on a VVR scaling
Are you scaling plate voltage and screen voltage but nothing else?
-
Yes, correct. As well as grid bias of course.
I’m taking the reverb transformer B+ off the non-scaled preamp Supply
-
I’m taking the reverb transformer B+ off the non-scaled preamp Supply
That could be a factor for your unique problem. That reverb driver is really just a small single-ended power amp. Fender wisely chose to power that driver from the screen node. I think it's a bad idea to power that driver from the preamp power supply. Your amp circuit has deviated significantly from the original AB763 design. I don't think the original design is your problem. I think your mods are the root of your problem.
-
Every amp that I've install a VVR into, had the preamp voltages increase when I dialed down voltages to the LTPI and power tube. Don't know that this would be a factor though?
With respect, Tubenit
-
Thanks for feedback. I can/will try running the transformer off a different point, but i don’t think that’s the issue (happy to be proved wrong though ;-) )
I’ll also measure the preamp voltages as well and report back.
At the end of the day, the input level and/or the gain of the reverb driver makes that stage clip at 6 or 7 on the volume control. Have a look at later fender model schematics and how fender altered the driver stage. I’m guessing to overcome this kind of issue (??)
If anyone has a deluxe reverb and a scope to hand. It would be interesting to see what happens to the reverb driver at high volume settings
-
Example of others experiencing this issue (or at least similar root cause)
http://ampgarage.com/forum/viewtopic.php?p=217066 (http://ampgarage.com/forum/viewtopic.php?p=217066)
-
Every amp that I've install a VVR into, had the preamp voltages increase when I dialed down voltages to the LTPI and power tube
I'm taking a slightly different approach to VVR. Rather than letting the cathode idle current on 6V6 tubes decrease as the plate voltage is decreased by VVR, I've scaled the bias so that the cathode idle current actually increases slightly as plate voltage os decreased.
see this thread for why http://el34world.com/Forum/index.php?topic=23048.0 (http://el34world.com/Forum/index.php?topic=23048.0)
I've attached
a) Schematic of single-channel AB763, no vibrato, with master volume and VVR scaling on 6V6 only
b) Schematic of VVR circuit I'm using in this amp
-
You still show your reverb driver powered from node B but you said you are powering it from the pream node which would be node D. Maybe you could try using that 22µF cap between the 200Ω and 10K dropping resistors? Might help. That would make it look a bit more like the DR circuit.
Your 47K MV doesn't look quite right. There should be a coupling cap between the pot and V4B plate. Where did you find a 47K-A pot?
Bottom line... If you're happy with it. don't change it. :wink:
-
Thanks slucky
You still show your reverb driver powered from node B
Duh, my error, thanks for pointing this out, now corrected in v2 schematic to show how it is actually wired (as you suggest as well)
Your 47K MV doesn't look quite right. There should be a coupling cap between the pot and V4B plate. Where did you find a 47K-A pot?
You are right again :icon_biggrin: my error. There is a 0.1 cap missing and the pot is a 50K audio taper. Corrected in V2 schematic!
-
I've come across this issue in a SF TR I was servicing a while ago.
The 470 ohm V3 cathode bias resistor has drifted up in value a bit, but not enough to stop its 12AT7 from overdissipating and failing.
I thought it best to replace the cathode resistor, and to get the tube dissipation a bit more reasonable, decided to use the BF value 2k2, and to bypass it just because.
There was horrid blatty blips of distortion when the signal level was sufficient to push V3 into grid conduction; it stopped when the cathode was unbypassed.
My suspicion is that it was actually astable oscillation due to positive feedback via the non-ideal 0V circuit common system used, as re-jigging the arrangements there a bit helped to stop it.
Even without that, a Fender BF reverb preamp channel overdrives horribly, due to V3 grid being the first point at which clipping occurs (using typical control settings and as the signal level is increased from 0).
I think that's why the SFs removed the cathode bypass when the master volume was introduced, ie with a master vol it's possible to hear the horrid preamp distortion, whereas previously, it was masked by power tube grid clipping.; the lowered cathode resistor value being in order to improve the stage's gain/power output with the cathode unbypassed.
I think that the best way forward is to use the BF 2k2//22uF, but add a 470 ohm resistor in series with the cap.
-
The cure for me was remove the 22uf cap. The result was the harsh distortion disappearing. fwiw I also intend to reduce the cathode resistor down to 1,500 ohms . . .
If the problem is grid current, and I believe this is what causes your problem, then reducing the cathode resistor will make the problem worse. As you decrease the value of the cathode resistor, it takes less signal on the grid to cause grid current. Also, if you keep the cathode resistor the same and go from unbypassed to bypassed, it takes less signal to cause grid current. So as far as unwanted grid current goes, a 470r fully bypassed cathode resistor is the worst case and a 2.2K unbypassed cathode resistor is the best case. The 470r case idles at a much higher current than the 2.2K case so that the 470r case is modulating more current giving it the ability to drive the tank harder, but that is not the concern here.
You could replace the 1M grid leak with a voltage divider so that it is like a set dwell control. I mean, those AB763's aren't hurting for reverb and cutting the signal at the driver grid a little would be acceptable for a DeLuxe unless you're playing the Clam Festival in Pismo.
-
Aren't some guys fixing this problem with grid stopper(s) or did i read those threads wrong?
-
Aren't some guys fixing this problem with grid stopper(s) or did i read those threads wrong?
When I first viewed the ampgarage thread, I just read the first few posts and decided it wasn't really about the same issue. Now that I read all three pages, you read it correctly and some people are fixing the problem with grid stoppers. My suggestion and pdf64's suggestion is also in there. It occurs to me that the grid stopper can alleviate the dry signal disruption, but still allow the driver to overdrive resulting is a less than desirable wet signal.
-
Thanks for the help and feedback.
You could replace the 1M grid leak with a voltage divider so that it is like a set dwell control
Before I removed the 22uf cap, I tried two things.
a. I added a 1meg resistor above the existing 1meg grid leak resistor, to half the input into the driver.
b. I added a resistor on the grid of the driver tube, a 68k I think
Neither of these had any effect, or none I could hear. Only when I lifted the 22uf did the issue seem to go away.
I really wish I had a scope, so I could see really what is going on.
-
If it's like the SF TR I mentioned earlier, a scope isn't super helpful.
I suggest to tinker with the circuit common implementation (grounding).
The screen grids and reverb driver should use the same chassis 0V connection as the power tubes, reservoir cap and HT CT.
If you've not already, have a read of http://www.valvewizard.co.uk/Grounding.pdf
-
ok, so I spent a few hours working on the amp today.
Firstly I completely altered the grounding. So that the reverb send tube's cathode resistor and the reverb transformer were earthed straight back to the star earthing point. This has made no change.
I then tried connecting the 22uf cap back in circuit but with a 470 ohm resistor in series, as was previously sugegsted. This still causes the distortion to happen, maybe not as bad, but it still happens. so I've removed it again.
If I disconnect the phono jack to the reverb tank and have volume on 10 i get no issues. As soon as I connect the reverb tank i get the distortion i described previously when the volume is above 7.
If i short out the reverb send as a test, I get the same distortion.
I'm guessing this is just how the design works on the blackface AB763 circuit? I'm presuming the reverb send tube is being overdriven at high volumes and in turn this is bleeding across the power rail onto the rest of the preamp causing the fizzy/farting distortion I'm hearing..
Perhaps for my application (slightly unusual?) where I'm looking to have the volume at 8 or 9 at low volumes, when i use a master volume and VVR on the output stage. perhaps I'd be better off using the reverb driver circuit as used on later twin reverbs?
such as
http://www.thevintagesound.com/ffg/schem/twin_reverb_sf_100_schem.jpg (http://www.thevintagesound.com/ffg/schem/twin_reverb_sf_100_schem.jpg)
and
http://www.thevintagesound.com/ffg/schem/twin_reverb_sf_135_schem.jpg (http://www.thevintagesound.com/ffg/schem/twin_reverb_sf_135_schem.jpg)
Or...would these give more gain and therefore the unwanted distortion again, when compared to the current setup which seems to now work ok? (a 2.2k cathode resistor, no bypass capacitor)
-
Both of those circuits would provide more drive current to the reverb tank.
-
Both of those circuits would provide more drive current to the reverb tank.
Thanks for that info.
So am I best sticking with what I've ended up with, or is another approach better? (I could put a reverb dwell pot on but would rather not).
-
your 50K MV pot is upsetting bias of the phase inverter, if it is indeed wired as shown. IOW, you're missing a cap from the MV wiper to the 220K to the grid of V6A.
see attached.
--pete
-
your 50K MV pot is upsetting bias of the phase inverter, if it is indeed wired as shown. IOW, you're missing a cap from the MV wiper to the 220K to the grid of V6A.
Take a look just north of your added cap and you'll see a .001µF coupling cap.
-
your 50K MV pot is upsetting bias of the phase inverter, if it is indeed wired as shown. IOW, you're missing a cap from the MV wiper to the 220K to the grid of V6A.
Take a look just north of your added cap and you'll see a .001µF coupling cap.
aha! i.c. now - getting old! :p
--pete
-
Hi,
yesterday I added the London Power SV1 power scaling kit with the RBX (bias supply) to my Hoffman AB763 one channel build without tremolo. So it's exactly the same amp but with the London Power LP-MV master volume... It's a pre phase inverter master volume with a dual 250k pot.
Now I have the same issue! I get a fizzy/crackly distortion from the reverb driver.
The reverb driver b+ is constantly at 393V...
-
Hi,
The Reverb driver distortion problem is well documented on the net over many years.
The trick is to make sure it is the " Reverb overdriving problem " first. If I get to this point I install a 560pF at greater than 1Kv across the anode to cathode pins of the driver tube, as in the Fender 75 amp.
Sometimes I do the Mesa boogie cap install with the cap across the driver transformer as well as the anode to cathode cap connection. Removing the decoupling cap from the cathode can help as well.
I usually find this distortion problem mostly in the Princeton reverb type circuits.
John
-
Hey all -
I've experienced this problem off and on with two different AB763 builds. It's really frustrating. Some observations:
-- Swapping driver tubes, reverb tanks, reverb transformers, and fresh bypass caps on driver/return tube never fixed it for me.
-- Removing the reverb driver's cathode bypass cap has always stopped the crackling distortion for me, however, it caused some feedback in the reverb that wasn't there prior. On my second AB763 build, I split the cathodes on the reverb return, and that all but stopped the feedback, so I ended up doing that to my first one as well.
-- Unfortunately, I feel like the reverb sounds better with the driver's cathode fully bypassed. But, I can't handle that crackling distortion, so I think it's a compromise I'm willing to make. But if anyone has any better suggestions or more recent findings to this problem, I'm all ears.
Also - I have NOT tried Merlin's suggestion of adding a damping resistor to the secondary and small cap on the primary (link attached) - has anyone tried this?
https://www.valvewizard.co.uk/reverbdriver.html
-
Also - I have NOT tried Merlin's suggestion of adding a damping resistor to the secondary and small cap on the primary (link attached) - has anyone tried this?
https://www.valvewizard.co.uk/reverbdriver.html
The fix isn't the resistor and cap (which are mostly protection mechanisms), but it is outlined elsewhere in that write-up. Create another power supply node that will give a B+ of about 350V. Change the cathode resistor to 1.5k. This will buy you a bit of headroom in the driver stage and should avoid the distortion you're hearing.
-
Would that be done on the primary side of the reverb transformer? Just a standard voltage divider? No filter cap needed I presume?
-
Would that be done on the primary side of the reverb transformer? Just a standard voltage divider? No filter cap needed I presume?
Yes. Primary side.
Just add another node (filter cap included). You can split it off from the same spot the transformer was originally connected. No need for a dedicated voltage divider.
-
Would that be done on the primary side of the reverb transformer? Just a standard voltage divider? No filter cap needed I presume?
Yes - the High Tension supply does go to the reverb transformer primary - so the new voltage divider and filter cap supply node needs to go between the RT primary and the existing power rail