AB763 deluxe reverb reissue conversion B+ and buzz problem

[Mrsedor]

Hi,
I started with a buggy / noisy PCB 65 Deluxe reverb reissue, chassis with all the transformers and choke and wanted to go for a fresh start, converting it to Eyelet, with better components and hopefully a better sound.

Long story short, I have 2 problems, one is the 50hz and 100hz noise (EU power here) that is a bit high, especially when V4 is in, with or without reverb, and not depending that much on the volume of either channel.
Plate voltage at V4a & b are good (170/180v).

The other is, with a probe connecting tool (1ohm in an intermediate socket), I cannot get the bias down to less than 29/30ma, which seems to high for a 6V6, I'd like it at 22ma but cannot reach it...
When powered without 6V6 but rectifier installed, I can reach -42V of negative bias maximum.

When I read the supply values, D+ looks fine at 280v, but C+ is at 340V instead of 325V and B+ is at 440V instead of 415V.

I followed precisely Rob robinette's layout (attached), with 2x100 ohms for Heater center tap, and with the Tick fix capacitor along the photocell.
I used orange Drop coupling capacitors, 5x FT 16uf, sprague atom and RN70 or RN65 dale metal film resistors.

I added a 370K 5w bleeder between first filter cap and ground. It drains all the power in 10sec.

Input jack are isolated from the chassis, with a preamp ground bus running behind them, connecting to chassis via a screw near the normal 1 input. Preamp ground is connected to this.

I run coax with shield grounded at this bus from 68ks inputs to grids and back to volumes for both channels.

RCA jacks for reverb send/return are not isolated and grounded to chassis, with reverb transformer ground and 220k input resistor connected at this ground point.

TRS jack for the footswitch is not isolated. Grounded directly at chassis. Should it be and grounded to the preamp bus ground? This is ground and reverb I/O ground is more distant than in the original, as the RCA plugs are under the chassis and not behind like the footswich plug.

Output jacks are not isolated, grounded directly at the chassis.

Power ground is near power transformer, on a different bolt than the earth, it connects HT centertap (red yellow), 6V6 cathodes, bias board and first stage of filtering.

My wiring does not look as good as certain builds I see on the web, but soldering is good and I tripled check the build with the schematic. Should I shorten the plates and cathodes wires to the minimum?

I had first a big buzz problem, and it got really better when I moved the V4 wiring from the BCD supply lines. It's still there a bit though. I wonder if i should drill another hole in the chassis and run it the way they do in the 64' reissue, meaning on the other side of the eyelet board. They look better twisted also. Should I go for it?

I have tried to swap the OT wires, because it doesn't squeal. These are not twisted. Should I do otherwise?

I will post detailed pictures.

On a side note, one 6V6 is a bit loose between the glass and and the black plastic base. I will order new ones.
Gently tapping with a stick V4 is very microphonic, is it normal?

Sorry, lots of questions, I'm a bit lost after long days of trying to figure it all...
Thanks in advance for your help!

[Mrsedor]

With more pictures :

[Mrsedor]

More pictures

[Mrsedor]

More pictures

[Mrsedor]

And also, reverb tank is connected, and it is not apparently creating a ground loop (isolated on one side of the tank).

Thanks :)

[Lectroid]

R.G. Keen's Troubleshooting page addresses a lot of your symptoms with concrete troubleshooting suggestions.

http://www.geofex.com/ampdbug/ampdebug.htm

[Mrsedor]

R.G. Keen's Troubleshooting page addresses a lot of your symptoms with concrete troubleshooting suggestions.

http://www.geofex.com/ampdbug/ampdebug.htm

Thanks a lot, I will look at it thoroughly :)

By the way, do you have any idea about why B+ might be a bit high? Or could it normal to be 440v instead of 420v ?

Thanks 🙏

[Lectroid]

First, please post a schematic of what you've actually built.  Without one we're all shooting in the dark.  Take one clear landscape shot and capture the entire board and chassis in one pic. 

If B+ is higher than expected, is is possible you have the voltage selector switch set incorrectly?  If not, is it really a problem?  Put in 6L6s and call it a Pro Reverb. 

Okay, what rectifier tube? New or used?  What's your screen voltage at "B" node?  Record voltages with all tubes in and the amp warm.

Yes, put in new 6V6s--they won 't hurt.  I have no experience wrapping control pot wires around others as a bundle, but a lot of builders do it.

Uncle Doug has a YouTube video about 60Hz vs. 120Hz hum (or 50/100Hz in your case) and how to tell the difference.  Also ValveWizard.com and Aiken Amps have grounding discussions as well as lots of other info. 

http://aikenamps.com/index.php/grounding

http://www.valvewizard.co.uk/Grounding.html

Bias: What's your max and min bias voltage at the 6V6 grids?  According to RobRob's calculator, 30mA is 90% of max dissipation.  The ValveWizard says 85% is fine for AB push-pull amps with fixed bias.  Whereas 22mA would put you at 66%, maybe a tad low?  Do you have the 1 ohm 1% resistors between each 6V6 cathode and ground? 

Fender biased low (60%) for tube longevity but if you're willing to replace tubes every few years then a hotter bias might be  acceptable to you.  How does it sound to you?  My AB763 rolls around 25mA (75%) on new JJs, and I'm quite happy with it.  I don't play out enough to worry about tube life.

The 470 ohm bias resistor controls the range of your bias voltages (assuming the stock Fender circuit.)   Has it drifted? 

[Mrsedor]

First, please post a schematic of what you've actually built.  Without one we're all shooting in the dark.  Take one clear landscape shot and capture the entire board and chassis in one pic. 

If B+ is higher than expected, is is possible you have the voltage selector switch set incorrectly?  If not, is it really a problem?  Put in 6L6s and call it a Pro Reverb. 

Thanks a lot !

The circuit I built is the one attached, I followed it precisely without any change but the 0.02uf across the optocell, and 0.05uf instead of 0.047uf. I used the layout attached in the first post.
It is possible I made a mistake, but I triple checked all the connexions already... I will check it again.

I used the transformer originally fitted in the 65 deluxe reverb reissue for EU, it was connected to pcb with the violet (live) and white (neutral). But they were other wires to the power transformer that were not connected to the circuit (attached to the PCB still).

[Mrsedor]

If B+ is higher than expected, is is possible you have the voltage selector switch set incorrectly?  If not, is it really a problem?  Put in 6L6s and call it a Pro Reverb. 

There was no voltage selector in the "original" pcb circuit.

Here is the diagram of the transformer fitted in international amplifiers, like mine.

I don't know if someone mess with this connexions before me in this amp.

Violet (live) and white (neutral) from the Mains Switch / fuse CP7 & CP8
Black and White , attached to the PCB but connected to nothing. I believe it was intended for 100V. CP1 & CP2

Black/Red CP3, connected to nothing.
Black/Yellow CP5 & Black/Green CP6, connected together, but to nothing else.
No CP4.
I took a picture of what it was before.

Here power is 230V.

[Mrsedor]

Okay, what rectifier tube? New or used?  What's your screen voltage at "B" node?  Record voltages with all tubes in and the amp warm.

The tube is a 5AR4 from electroharmonix, that was before in this amp when it was a Drri.
I will record all the voltage as soon as i can, probably on Tuesday, and report back.

Thanks a lot for your help !

[Lectroid]

I wasn't asking for all voltages.  My comment was just that once the amp is running, always test with all tubes in, ie, under full load. 

A schematic would be more valuable than all voltages at this point.

[Mrsedor]

I wasn't asking for all voltages.  My comment was just that once the amp is running, always test with all tubes in, ie, under full load. 

A schematic would be more valuable than all voltages at this point.

The schematic is this one, attached to the post.
Thanks :)

[Mrsedor]

Take one clear landscape shot and capture the entire board and chassis in one pic. 
 

Is this picture (attached to this post) good enough?

[Lectroid]

The picture is fine, the schematic is the standard.  All good. Your layout is good and your soldering and lead dress are probably better than mine.

There was no voltage selector in the "original" pcb circuit...
Violet (live) and white (neutral) from the Mains Switch / fuse CP7 & CP8 ... I took a picture of what it was before.
Here power is 230V.

Okay. 
1. What is your actual wall voltage at any wall socket, measured by voltmeter? 
2. What is the HT voltage going to pins 4 and 6 of the 5AR4?  If your DVM doesn't have a range of 0-1000V, then just measure from pin 4 or pin 6 to ground, that'll give half the value, then double it.  (Fender's schematic suggests to me that their transformer puts out around 320-0-320V P-P, ca. 650VAC. )
3.  What are your 'A', 'B', 'C', and 'D' node voltages.
4.  What is the total voltage put out by your bias circuit?
5.  What is the maximum and the minimum bias voltages you can get?  In other words, the bias voltage at the 6V6 grids with the bias pot turned all the way up and then all the way down.

I've attached an export transformer wiring scheme from a Fender DRRI 65 manual, see below.  It doesn't match what you described, but hopefully it'll help to confirm/disprove your conclusions about which windings you're using.

Honestly, although your voltages are slightly higher than nominal it's hard to see them as a 'problem.'  The 230V version (of the DRRI transformer) shows four other secondaries from the DRRI that an AB763 circuit doesn't use.

On one hand that would explain why your AB763 version has 'extra' power available to it.  It also makes me wonder about how you wired up your bias circuit.  In other words, the 'real' AB763 transformer had a dedicated bias tap.  Where do you tap power for your bias circuit?

The hum issues are harder to diagnose from a distance.  I still recommend Uncle Doug's video; he'll point you in the right direction.  Keen's article(s) have a lot of things to try.  Report back on what you find.

Good luck!

[Mrsedor]

1. What is your actual wall voltage at any wall socket, measured by voltmeter? 
2. What is the HT voltage going to pins 4 and 6 of the 5AR4?  If your DVM doesn't have a range of 0-1000V, then just measure from pin 4 or pin 6 to ground, that'll give half the value, then double it.  (Fender's schematic suggests to me that their transformer puts out around 320-0-320V P-P, ca. 650VAC. )
3.  What are your 'A', 'B', 'C', and 'D' node voltages.
4.  What is the total voltage put out by your bias circuit?
5.  What is the maximum and the minimum bias voltages you can get?  In other words, the bias voltage at the 6V6 grids with the bias pot turned all the way up and then all the way down.

Hi,
Thanks for your help!

Here are today's reading:

With all tubes in the sockets.

1. Wall socket : 238V

2. 5AR4 UpP4 Ground : 343V   UP6 Ground : 343V

3. A (standby open, off) 461V    weird?
    A (Standby on) 420V
    B 420V in all 3 points in the circuit
    C 333V
    D 272V

4. With the 2x 6V6 plugged:
    at both grids (pin5) -37.1V
    at 220k resistors junction -45,2V

[sluckey]

3. A (standby open, off) 461V    weird?
    A (Standby on) 420V
    B 420V in all 3 points in the circuit
    C 333V
    D 272V

These look very normal. "A" high voltage is normal when STBY is open.

4. With the 2x 6V6 plugged:
    at both grids (pin5) -37.1V
    at 220k resistors junction -45,2V

The voltage on the grids should be the same as the voltage at the 220K junction. Recheck that.

OK, you have about 420V on the plates of the 6V6s and about -37V on the grids. That sounds normal to me. My DR has 423V and -36V and that gives me 24mA cathode current using 1Ω current sense resistors on the cathodes. I  think something is not right with your bias probe. I suggest you get two 1Ω 1W 1% resistors and connect them between cathode (pin 8) and ground and use those to measure your bias current.

[Mrsedor]

Without the 6V6 in the sockets:

Minimum bias voltage at the grid (p5) : -36,2V
Maximum voltage: -18V

Is it correct?

[pdf64]

With fixed bias amps, the typical 10M input resistance loading effect of most DMMs causes a small (2%) volt drop across the typical 220k grid leaks.
So -50V at the bias supply output will measure -49V at the control grid socket terminal.

Dropping 8V across 220k implies a meter loading resistance of 1M.

I suggest to measure the voltage from one end of a 220k grid leak to the other.
 

[sluckey]

Without the 6V6 in the sockets:

Minimum bias voltage at the grid (p5) : -36,2V
Maximum voltage: -18V

Is it correct?

I don't like that range. It means that the bias pot needs to be set to one extreme to get a safe bias. But you say you have -45V at the junction of the 220K resistors. You should also have that same -45V at the grids in a properly working amp. Find out why that is not so. Leave the 6V6s unplugged until you resolve this. About the only thing that would cause this difference in a properly wired amp would be leaky coupling caps between the PI and the 6V6 grids. Temporarily disconnect both of those .1µF couplers and see if you now have the same voltage on the grids and the junction of the 220Ks.

BTW, meter are you using?

[Mrsedor]

At the jonction between the 2x 220k and the wire from the diode I have :
Minimum -45v
Maximum -22V

At the jonction between 220k and .1 coupling cap I have:
Minimum -37V
Maximum -18V

At Pin 1 and at Pin 5, after the 1.5K, I have the same:
Minimum -37V
Maximum -18V

[pdf64]

Right, but that’s negligible.

[Mrsedor]

Right, but that’s negligible.

At the jonction between the 2x 220k and the wire from the diode I have :
Minimum -45v
Maximum -22V

At the jonction between 220k and .1 coupling cap I have:
Minimum -37V
Maximum -18V

At Pin 1 and at Pin 5, after the 1.5K, I have the same:
Minimum -37V
Maximum -18V

[pdf64]

I suggest to measure the voltage from one end of a 220k grid leak to the other.

[Mrsedor]

With fixed bias amps, the typical 10M input resistance loading effect of most DMMs causes a small (2%) volt drop across the typical 220k grid leaks.
So -50V at the bias supply output will measure -49V at the control grid socket terminal.

Dropping 8V across 220k implies a meter loading resistance of 1M.

I suggest to measure the voltage from one end of a 220k grid leak to the other.

I mesured both 220k. For each mesurement, one probe at the 2x 220k jonction and the other at the 220k to wire jonction).

0V dc

[Mrsedor]

I suggest to measure the voltage from one end of a 220k grid leak to the other.

From the 220K to wire jonction of each 220K (not at their jonction) I also mesure 0V dc

[Mrsedor]

The resistors are within specs when mesured inside the circuit (218k)

[Mrsedor]

BTW, meter are you using?

I'm using a simple DMM, Mastech MAS830

[Mrsedor]

One thing I don't understand is:

I have -45v between ground and 220Ks jonction
I have -37v between ground and the grid P5
I have 0v between 220Ks jonction and the grid P5...

 

[Mrsedor]

About the only thing that would cause this difference in a properly wired amp would be leaky coupling caps between the PI and the 6V6 grids. Temporarily disconnect both of those .1µF couplers and see if you now have the same voltage on the grids and the junction of the 220Ks.

I disconnected both .1 couplers, even the wires going to the grid stoppers.
Still the same, at the jonction of 220k, I have 45v, then 37V after each 220K.
So 8V is dropped across them... not normal, right?

[Mrsedor]

It is always with 6V6 out of the sockets.

[sluckey]

I disconnected both .1 couplers, even the wires going to the grid stoppers.
Still the same, at the jonction of 220k, I have 45v, then 37V after each 220K.
So 8V is dropped across them... not normal, right?

NOT NORMAL! So, you had the caps and the wires disconnected. That means the only thing in that eyelet is one lead of the 220K. That's an open circuit. Current can't flow and there will be no voltage dropped across the resistor. UNLESS...

There is continuity between that eyelet and chassis. Maybe something under the board is touching the eyelet, such as a screw? Or, the board is conductive, which happens a lot with the type board you are using. Where did you get that board?

Try this. Lift the ends of the 220K resistor from the board, but leave the junction connected. Now measure voltage on each end of the two 220K resistors. I bet they are all the same now.

[Mrsedor]

There is continuity between that eyelet and chassis. Maybe something under the board is touching the eyelet, such as a screw? Or, the board is conductive, which happens a lot with the type board you are using. Where did you get that board?

Try this. Lift the ends of the 220K resistor from the board, but leave the junction connected. Now measure voltage on each end of the two 220K resistors. I bet they are all the same now.

I got that eyelet board from Banzaii music in Germany :
http://www.banzaimusic.com/eyelet-board-blackface-reverb-ab763.html

I will test the continuity with the chassis and disconnect like you suggest.

[Mrsedor]

By the way I had a look at the DRRI bias schematic. My power transformer comes from it, so maybe it doesn't deliver the right voltage to the bias circuit.

In this bias circuit, the 470Ohm is a 22Ohm, and 100uf replace the 25uf I'm using.

[Mrsedor]

Try this. Lift the ends of the 220K resistor from the board, but leave the junction connected. Now measure voltage on each end of the two 220K resistors. I bet they are all the same now.

OK, I did it, disconnected the 220k from the circuit. Wire from the bias pot soldered to the 2x 220k, in the air. One leg of each 220k not connected to anything but the meter probe.

Nothing is connected to the eyelet board.

Still the same voltages...
 

[Mrsedor]

OK I suspected the DMM.
I tried another one and.... NOW I get the same voltages!

I will solder it back to how it was before and mesure with the new DMM at the grid.

[Mrsedor]

So... everything is soldered back where it was...
With the other DMM I read -45V at Grid...

It was a DMM problem after all 

So now at the grids I can read from -45V to -22.

Is that ok?

[sluckey]

So now at the grids I can read from -45V to -22.

Is that ok?

very good.

[Mrsedor]

very good.

Thanks! I think I can get the right bias now :)

But, regarding my hum problem :

Amp is not humming when only the 6V6s are plugged.

PI plugged, a bit of hum, not much but still noticeable.
V1 plugged, the same.
The hum is not increasing with the volume.

V4 plugged. More hum.
V2 plugged, the same.
Hum still not increasing with the volume.

I suspect something is picked around V4 and V6.

I have a suspect. As I converted this DR from a DRRI, the access from the doghouse to the chassis is a hole located near the tube connexions. In the original and the 64 reissue, it looks like the A/B/C wires are coming from the other side of the circuit, with a more direct access to their eyelet connexions.

In my circuit, now those wire are running across the board as shown in the picture attached. At first they were really close to the tube wires and moving those cables with a stick was changing the noise quite a bit.

Do you think I should drill a hole to route the ABC wires the same way as the original, thus keeping those wire as far from the board as possible? It's a bit of work, but I will do it if you think it might improve this hum.

Thanks a lot for your help!

[Mrsedor]

I did drill the chassis and re route those wires like they do in the 64 DRRI.

It didn't help with the hum....

[Mrsedor]

Could this hum be related to a bad Heater wiring?

The heater wires are close to the other wires, like in the picture.

[Mrsedor]

very good.

As for the bias, I'm scratching my head.

With this bias probe, I'm getting minimum 29mA, which is too much...
I will try to install those 1 ohm resistors.

It's between cathode and power ground, right?

[sluckey]

As for the bias, I'm scratching my head.

With this bias probe, I'm getting minimum 29mA, which is too much...
I will try to install those 1 ohm resistors.

It's between cathode and power ground, right?

If you are getting 29mA when you have -45VDC on pin 5 I would suspect your bias probe is lying to you. Install the 1Ω resistors between pin 8 and ground.

[Mrsedor]

If you are getting 29mA when you have -45VDC on pin 5 I would suspect your bias probe is lying to you. Install the 1Ω resistors between pin 8 and ground.

I just installed the 2x 1ohm 5w resistors and the reading is : 29mV across the first, and 27.5mV across the second. This is the minimum I can get.

Not good, right?

[Mrsedor]

Regarding the Hum, it is humming with the PI and 6V6 installed.

It is low 50hz hum, not changing with the volume (when the other tubes are installed).

I have several grounds:

AC outlet ground, alone, bolted to chassis.
Output jacks ground, bolted to chassis
Pedal trs jack ground, bolted to chassis
RCA send/return reverb ground, bolted to chassis.
Power ground, with HTCT (red/yellow, bias ground, 6V6 cathodes ground, first filtering stage ground). Bolted near PT.
Preamp ground, on a bus wire running from PI ground (at 47 ohm) to jack, getting all the other small signal grounds, bolted near the jack input. All 4 jack inputs are isolated.

Do you think it could be a ground loop? I though I respected the layout, but I could have made a mistake...

What do you think?

Thanks a lot,
 

[Lectroid]

By the way I had a look at the DRRI bias schematic. My power transformer comes from it, so maybe it doesn't deliver the right voltage to the bias circuit.

In this bias circuit, the 470Ohm is a 22Ohm, and 100uf replace the 25uf I'm using.

A dedicated bias tap should be in the 50-60V range.  Why don't you measure the (AC) voltage coming into your bias dropping circuit?  What is that voltage?  Shouldn't be too hard to measure if you have clip probes.  If it's much higher maybe that's one problem.

Also, that 220 ohm resistor looks low to me but I defer to sluckey and pdf64.

[Mrsedor]

A dedicated bias tap should be in the 50-60V range.  Why don't you measure the (AC) voltage coming into your bias dropping circuit?  What is that voltage?  Shouldn't be too hard to measure if you have clip probes.  If it's much higher maybe that's one problem.

Also, that 220 ohm resistor looks low to me but I defer to sluckey and pdf64.

At the PT tap red/blue and 470 Ohm resistor jonction I read 37 VAC to Power Ground.
This could be the problem?

Thanks a lot

[Mrsedor]

I read on the normal Deluxe Reverb PT specs that this is supposed to be 50VAC.

So either mine has a problem or the DRRI has a different PT, which would kinda explain the different bias circuit?

Do anyone has ever encountered this issue?

[sluckey]

The DRRI uses a different PT than used by an original AB763 DR. So, replace your 470Ω with a 220Ω like the DRRI uses. You'll get a little more -VDC.

But, you already said, "So now at the grids I can read from -45V to -22". That's a perfectly good bias voltage range for 6V6s that will likely be set to about -36V when biased properly. There is no issue here.

BTW, my DRRI PT puts out 38VAC on the bias tap and I used a 470Ω with no issues. And my bias pot is set for -36V on the grids and that gives 24mV on the cathode 1Ω resistor.

You need 1Ω 1% 1W cathode resistors for accurate mV readings on the cathode. Your 1Ω 5W resistors are probably only 5% tolerance. Which meter are you using now? Maybe it doesn't read dc mV accurately.

Try this... adjust the bias pot to provide -37V at the junction of the 220K grid resistors. This is according to the DRRI schematic. Do the 6V6s redplate? Ignoring the hum, how does the amp sound?

[Mrsedor]

Try this... adjust the bias pot to provide -37V at the junction of the 220K grid resistors. This is according to the DRRI schematic. Do the 6V6s redplate? Ignoring the hum, how does the amp sound?

Hello,
Sorry about the delayed reply, I had to manage health issues and just got home now.

So, I progressively provided -37V at the junction of the 220k. The 6V6 both redplated, so I switched the standby off and reverted to -45V.

This is not normal, right?

Thanks a lot.