Bias problem!

[thomasdj]

Hi!

I am building the 3ch dual rectifier and I am almost finished...I think!

Well..i have made a few modifications to the schematic..I added individual bias pots to each power tube.

I 've got two problems.

I got signal through each channel and so I assume that the preamps are working properly. However I've got a nasty high freqency sound coming out of the speaker. Another problem is that I only get low power even when turned at high gain and master.

I run the amp with 4xEL34 so i adjusted the bias negativ voltage to ca. -40v at each power tube at Grid 1 (pin 5). But when measuring cahtode voltage across 1 ohm inserted cathode resistor I only measure 10-12 mV!!! ..what is wrong with my circuit??

Is it the 4x220K that should be lower??

Thanks

Thomas

[scole]

hey thom, the noise sounds like negative feedback wich you can solve by switching the ot leads to the power tubes. the other problem im not expert enough to help.

[MWJB]

What kind of plate voltages do you have on the EL34's?

[thomasdj]

What kind of plate voltages do you have on the EL34's?

The plate voltage is 475..pretty close to the schematic..

Btw..the tubes get really hot! ..so do the PT ..when I tap the input jack tip ..the tubes "blink" and the OT gnits at the plate voltage:s

[MWJB]

At 475 I'd suggest that you want more than -40v at the grids. Do you have 1x 1ohm resistor at each tube cathode? Are you sure that you are not measuring for 2 tubes & have the decimal point in the wrong place?

[thomasdj]

At 475 I'd suggest that you want more than -40v at the grids. Do you have 1x 1ohm resistor at each tube cathode? Are you sure that you are not measuring for 2 tubes & have the decimal point in the wrong place?

I attached my bias measure setup..
Don't think that there is anything wrong with it..

But as I mentioned, the tubes get really hot! and when I tap the input jack tip the tubes "blink" and the OT sparkles at the plate voltage:s

[thomasdj]

hey thom, the noise sounds like negative feedback wich you can solve by switching the ot leads to the power tubes. the other problem im not expert enough to help.

Thanks Scole:)

But I have to solve the bias problem before doing anything else:)

[Bruce / Mission Amps]

Thanks Scole:)

But I have to solve the bias problem before doing anything else:)
Sounds more like ultrasonic oscillations to me.
Do you have an O'scope?
Disconnect the NFB completely and see if that helps... also if you have a presence control wired in there, check your lead dress with respect to the phase inverter... move the wire around and try grounding it somewhere else.. etc.
I think you have a wire dressing problem and a higher gain circuit is leaking signal through capacitively coupled wires back to a prior stage at some very high frequency.
Many the times it is right at the phase inverter but it can be anywhere with respect to grid lead wires.

[thomasdj]

Sounds more like ultrasonic oscillations to me.
Do you have an O'scope?
Disconnect the NFB completely and see if that helps... also if you have a presence control wired in there, check your lead dress with respect to the phase inverter... move the wire around and try grounding it somewhere else.. etc.
I think you have a wire dressing problem and a higher gain circuit is leaking signal through capacitively coupled wires back to a prior stage at some very high frequency.
Many the times it is right at the phase inverter but it can be anywhere with respect to grid lead wires.

Thanks Bruce..I'll try to do that right now! and see if it helps:)

[thomasdj]

Sounds more like ultrasonic oscillations to me.
Do you have an O'scope?
Disconnect the NFB completely and see if that helps... also if you have a presence control wired in there, check your lead dress with respect to the phase inverter... move the wire around and try grounding it somewhere else.. etc.
I think you have a wire dressing problem and a higher gain circuit is leaking signal through capacitively coupled wires back to a prior stage at some very high frequency.
Many the times it is right at the phase inverter but it can be anywhere with respect to grid lead wires.

Well...Don't know if it did something...but found a new thing..! when I touch the PT on the iron..some of the noise disappear! ..and the sparkling also disappear when tapping the jack tip and touching the PT iron simultaneously..

No..don't have a O'scope

[Steve Conner]

Sparkling?! Are you seeing sparks from the wires on your output transformer? If so, you need to stop right now before you wreck it. This sounds like ultrasonic oscillations, sure enough.

Try running the amp with no tubes fitted except the rectifiers and power tubes. Then fit the PI tube and try again. Then add the next tube back in the signal path, try again, and so on.

It can help to measure the plate or cathode current draw. If the amp is working properly, the idle current should adjust easily with the bias pot. You should be able to set it to 84mA or whatever, and it will stay there, only increasing when you play loud. If there are oscillations, the adjustment will be nearly impossible, and the current will jump to some high value when the oscillations start. Try all the controls and see if they affect the current draw. Particularly the volume controls and the presence control. Even moving your hand around the chassis may affect it.

Did you try reversing the plate connections to your OT? You might have the phasing wrong, so your negative feedback is actually positive.

Also, you might need a different value of compensating capacitor than the Dual Recto schematic shows. The compensating capacitor is usually hooked between the PI plates. It is the 75pF cap in your schematic. The required value depends on the make and model of OT you're using. If the amp oscillates with only the PI tube, rectifiers and power tubes fitted, and you're sure you have the OT plate connections right, then try a bigger value for this cap, maybe 100 or 150pF.

This is just a brain dump: I'm not personally familiar with the Dual Recto, but I've debugged ultrasonic oscillations before. The amp would draw 7 times the normal idle current and redplate as soon as the dirty channel was engaged and the master volume was turned above 10 o'clock. The cure was a tube shield on the first preamp tube.

[thomasdj]

Sparkling?! Are you seeing sparks from the wires on your output transformer? If so, you need to stop right now before you wreck it. This sounds like ultrasonic oscillations, sure enough.

Try running the amp with no tubes fitted except the rectifiers and power tubes. Then fit the PI tube and try again. Then add the next tube back in the signal path, try again, and so on.

It can help to measure the plate or cathode current draw. If the amp is working properly, the idle current should adjust easily with the bias pot. You should be able to set it to 84mA or whatever, and it will stay there, only increasing when you play loud. If there are oscillations, the adjustment will be nearly impossible, and the current will jump to some high value when the oscillations start. Try all the controls and see if they affect the current draw. Particularly the volume controls and the presence control. Even moving your hand around the chassis may affect it.

Did you try reversing the plate connections to your OT? You might have the phasing wrong, so your negative feedback is actually positive.

Also, you might need a different value of compensating capacitor than the Dual Recto schematic shows. The compensating capacitor is usually hooked between the PI plates. It is the 75pF cap in your schematic. The required value depends on the make and model of OT you're using. If the amp oscillates with only the PI tube, rectifiers and power tubes fitted, and you're sure you have the OT plate connections right, then try a bigger value for this cap, maybe 100 or 150pF.

This is just a brain dump: I'm not personally familiar with the Dual Recto, but I've debugged ultrasonic oscillations before. The amp would draw 7 times the normal idle current and redplate as soon as the dirty channel was engaged and the master volume was turned above 10 o'clock. The cure was a tube shield on the first preamp tube.

Many thanks Steve!

I really think this is what happens! ..it all makes sence.. I'll try the amp without the preamp tubes and so on...

I didn't swap the plate connections but I'll do it now! but what will you suggest me to do first?? swap the plate connections or pull out the tubes??

Btw...here is a link to the OT I use..
http://www.welter-electronic.de/bauteile/trafos/uebertrager/standard/standard_uebertrager.htm

number Ü23 Fender Bassman 100

sorry it's in german..

[thomasdj]

Sparkling?! Are you seeing sparks from the wires on your output transformer? If so, you need to stop right now before you wreck it. This sounds like ultrasonic oscillations, sure enough.

Try running the amp with no tubes fitted except the rectifiers and power tubes. Then fit the PI tube and try again. Then add the next tube back in the signal path, try again, and so on.

It can help to measure the plate or cathode current draw. If the amp is working properly, the idle current should adjust easily with the bias pot. You should be able to set it to 84mA or whatever, and it will stay there, only increasing when you play loud. If there are oscillations, the adjustment will be nearly impossible, and the current will jump to some high value when the oscillations start. Try all the controls and see if they affect the current draw. Particularly the volume controls and the presence control. Even moving your hand around the chassis may affect it.

Did you try reversing the plate connections to your OT? You might have the phasing wrong, so your negative feedback is actually positive.

Also, you might need a different value of compensating capacitor than the Dual Recto schematic shows. The compensating capacitor is usually hooked between the PI plates. It is the 75pF cap in your schematic. The required value depends on the make and model of OT you're using. If the amp oscillates with only the PI tube, rectifiers and power tubes fitted, and you're sure you have the OT plate connections right, then try a bigger value for this cap, maybe 100 or 150pF.

This is just a brain dump: I'm not personally familiar with the Dual Recto, but I've debugged ultrasonic oscillations before. The amp would draw 7 times the normal idle current and redplate as soon as the dirty channel was engaged and the master volume was turned above 10 o'clock. The cure was a tube shield on the first preamp tube.

Hi again Steve

I am pretty new in measuring mV but anyway...I switched the multimeter to 200mV ..the display then shows 00.0

I measured the cathode current on each power tube with the PI plugged in. The meter shows:

00.8
04.1
01.9
00.2

The bias voltage is -40v on each tube. If i turn up the negative voltage the tube grids glows blue!

However..The oscillations starts when V1 is plugged in and the gain or master is turned slightly..

What to do now??

[Steve Conner]

Well, those mV readings suggest you just are biased too cold. You need to adjust the negative voltage to be less than -40V, say -38, -35, or whatever, and as you do this the mV readings will all increase. Adjust it so that you get roughly 30mV across each cathode resistor.

Blue glow inside the tubes is normal.

If the oscillation only starts with V1 plugged in, then it sounds like a layout problem, that might need shielding or changes to your wiring layout to solve it. Also, it suggests that the plate wires on your OT were the right way round to start with, so probably no need to bother changing them.

Does the oscillation happen on all three channels, or only on the highest gain one? (red channel or whatever Mesa Boogers calls it?) Can you post photos of your chassis so we can see the layout?

In particular, if V1 is near to the power tubes, then it might need a shield. For a test, try wrapping V1 in tin foil connected to the chassis, and see if this affects the oscillations.

[thomasdj]

Well, those mV readings suggest you just are biased too cold. You need to adjust the negative voltage to be less than -40V, say -38, -35, or whatever, and as you do this the mV readings will all increase. Adjust it so that you get roughly 30mV across each cathode resistor.

Blue glow inside the tubes is normal.

If the oscillation only starts with V1 plugged in, then it sounds like a layout problem, that might need shielding or changes to your wiring layout to solve it. Also, it suggests that the plate wires on your OT were the right way round to start with, so probably no need to bother changing them.

Does the oscillation happen on all three channels, or only on the highest gain one? (red channel or whatever Mesa Boogers calls it?) Can you post photos of your chassis so we can see the layout?

In particular, if V1 is near to the power tubes, then it might need a shield. For a test, try wrapping V1 in tin foil connected to the chassis, and see if this affects the oscillations.

Hi Steve thank you!

I hava only tried the clean channel..since my OT spark from the primary I didn't want to push it..you know..

You wrote that blue glow in tubes are normal...is it normal that the grids are blue?? because when I turn down the bias voltage the grids turn blue..is this normal??

I have some pics i can send you..can't post them since they are to big
I know I have a layout problem..see pics:) ..don't get confused..all the "loose" wires are related to the switching circuit. The heavy red and heavy blac wire are both high power wires.. The heavy blue wires are connected to the preamp tubes..Underneath the PCB's are the power section wired PTP.. Hope it help's

[Steve Conner]

No, I'm not confused :eek: but if you wired an amp like that, you probably are! ;)

Where is the input jack? Is that it in the top right hand corner of the picture?

Where does the wire from the input jack run? Is it the thick blue wire that I see leaving the back of the jack and snaking in all directions through the chassis?

Is it shielded cable? If not, that's probably why the amp is oscillating.

Also, are the speakers connected correctly? If not, that might explain the sparking and quiet sound.

Yes, seeing a blue glow inside the electrodes while tubes are passing current seems to be normal. Clouds of glowing blue gas inside the tube are bad, though, and if anything else glows red except the heater, that's bad too.

[Chuck H]

Whoah... I don't want to upset anyone but...

Theres no possible way that amp could be stable. You really need to build a whole new board and wiring layout that has the circuit running in line with the tubes and ALOT less crossover in the stages. I didn't examine it long, but I can't imagine the grounding scheme is helping. And that can be critical in a high gainer.

Maybe you have to rebuild the board and rewire it, but you already have the chassis, pots, tubes and transformers in place. And you know what the likely culprit is, and it's fixable. But you absolutely can't cut any time corners when it comes to laying out a high gain amp.

I didn't even know if I should send this post. But Steve was very diplomatic about it, so... You really need to do a new layout.

no offence, Just want to help.

Chuck

[thomasdj]

Hi friends:)

Yep, I know and thanks for all of your help. This is my first build of a high gain amp..and PCB layout amp..

When I started this project I only had 4" x 6" PCB boards available which is why it was difficult to place components and so on..but now I got no limits! I got double side PCB's!

I know when making PCB amps the signal should be on one side of the board..and ground on the opposite side..I keep the power supply circuit as is..which is stargrounded to chassis underneath the PCB's (Left side in picture). But what about the preamp circuit? ..the board will be at least 16" long and 4" wide.. how should the different stages be grounded?? If this was a PTP build I would just go for the starground methode (near the OT). And what about the switching circuit?? where should it be placed?

Untill now I have only placed the components in my new PCB layout..from input signal to PI.

Are there anything else I should know when making PCB amps? If I have to route the signal on both sides of the PCB are there any limitations? like distance for instance??

Thank you so much!

Thomas

btw..sorry if my spelling is poor..I am from Denmark in Europe where we talk like #%¤&"""#### and grunts all the time;):D

[Steve Conner]

The main problem in tube amp layout is capacitive coupling. If you put a wire connected to a later stage near a wire connected to an earlier stage, the stray capacitance between those wires will let signal feed back, and make the amp oscillate.

Even a tiny stray capacitance can ruin your day. As an example, the worst thing you could do in the world of layout would be to run the input jack wire near one of the power tube plate wires. A high gain amp could get unstable if the plate wire came within about, say, 6 inches of the input jack wire.

Generally, we avoid that kind of screw-up by laying out all the components in a line that matches the signal flow. So the signal travels from one end of the chassis to the other, getting bigger as it goes, and never loops back on itself. Wires associated with later stages never need to cross wires associated with earlier ones, and the wires can all be short.

As you get more experience, you'll know what you can get away with. For example, you can get wires from different stages as close together as you like, IF you shield them. You often have to cross wires if you want an arrangement of front panel pots that doesn't match the signal flow.

For your amp, I'd recommend replacing the wire from the input jack to the first stage grid with shielded cable. If that doesn't work, start replacing all the other long runs of signal wire one by one, starting with the ones associated with V1.

BTW, I have friends in Denmark, and they have better coffee than Scotland at least :p

[thomasdj]

The main problem in tube amp layout is capacitive coupling. If you put a wire connected to a later stage near a wire connected to an earlier stage, the stray capacitance between those wires will let signal feed back, and make the amp oscillate.

Even a tiny stray capacitance can ruin your day. As an example, the worst thing you could do in the world of layout would be to run the input jack wire near one of the power tube plate wires. A high gain amp could get unstable if the plate wire came within about, say, 6 inches of the input jack wire.

Generally, we avoid that kind of screw-up by laying out all the components in a line that matches the signal flow. So the signal travels from one end of the chassis to the other, getting bigger as it goes, and never loops back on itself. Wires associated with later stages never need to cross wires associated with earlier ones, and the wires can all be short.

As you get more experience, you'll know what you can get away with. For example, you can get wires from different stages as close together as you like, IF you shield them. You often have to cross wires if you want an arrangement of front panel pots that doesn't match the signal flow.

For your amp, I'd recommend replacing the wire from the input jack to the first stage grid with shielded cable. If that doesn't work, start replacing all the other long runs of signal wire one by one, starting with the ones associated with V1.

BTW, I have friends in Denmark, and they have better coffee than Scotland at least :p

:D Yep Steve...we have good coffee! I do use shielded cable at the input jack (it's the grey wire upper right on picture..that goes straight to V1).

But what about the switching circuit and the ground issue? could it be routed at the same side of the PCB..the one facing down? should I starground each stage or can I make one big ground plane of the entire preamp inclusive the switching circuit? How long shall the traces between components be..to avoid any possibility for hum, oscillation etc..?

Thanks

Thomas

btw..Ok, we have better coffee in Denmark;) ..however I would love to try some of your fishing lakes in Scotland! ..since the danish fishing environment is totaly dead:mad:

[Enzo]

Steve knows a lot more about board layouts than I do, but a guitar amp is not a critical application. You have to pay attention, but it is not a high speed digital circuit nor does it run at RF freqs. From what you have posted it sounds like you are concerned over ground planes on the pc borad and such. You don't have to worry that a couple traces become a transmission line reflecting the signal back at you. At least not as a major concern.

Plenty of amps built on single sided boards, and I would not be so sure I would even want a ground plane covering one side of a two sider. It becomes a giant capacitor.

You can easily star the grounds on a pcb, ground traces tunning all down to the one point where they join. I see amps that way all the time, as well as video monitor chassis, and so on.

Look closely at a commercial amp board. Two parallel traces can couple as Steve mentioned. But you can also run a third parallel trace between them and have it part of ground (just connected at one end) and it serves to shield them one from another, or at least decouple them. First time I saw traces to nowhere, I was confused.

In my head, ground issues are more about shared current paths than where they are physically. In other words, if you have a relay switching parts in the first stage, you don't want the coil current using the same ground return as the signal circuits, otherwise the coil supply ripple current will be added to your signal. Doesn't matter that the relay is right next to the amp stage, it matters that they share copper.

I hope I haven't missed your point entirely. While we want to keep RF noise out of our amos, we don't need to build then like they were radios. I see ground planes as more a high freq systems element, not a hum abatement technique.

[thomasdj]

Steve knows a lot more about board layouts than I do, but a guitar amp is not a critical application. You have to pay attention, but it is not a high speed digital circuit nor does it run at RF freqs. From what you have posted it sounds like you are concerned over ground planes on the pc borad and such. You don't have to worry that a couple traces become a transmission line reflecting the signal back at you. At least not as a major concern.

Plenty of amps built on single sided boards, and I would not be so sure I would even want a ground plane covering one side of a two sider. It becomes a giant capacitor.

You can easily star the grounds on a pcb, ground traces tunning all down to the one point where they join. I see amps that way all the time, as well as video monitor chassis, and so on.

Look closely at a commercial amp board. Two parallel traces can couple as Steve mentioned. But you can also run a third parallel trace between them and have it part of ground (just connected at one end) and it serves to shield them one from another, or at least decouple them. First time I saw traces to nowhere, I was confused.

In my head, ground issues are more about shared current paths than where they are physically. In other words, if you have a relay switching parts in the first stage, you don't want the coil current using the same ground return as the signal circuits, otherwise the coil supply ripple current will be added to your signal. Doesn't matter that the relay is right next to the amp stage, it matters that they share copper.

I hope I haven't missed your point entirely. While we want to keep RF noise out of our amos, we don't need to build then like they were radios. I see ground planes as more a high freq systems element, not a hum abatement technique.


Thanks Enzo!

A lot of things make sence for me now!:) ..especialy the thing with traces going nowhere! ..thanks for the tip with the switching circuit! I'll make ground points rather than one big ground plane! ..I am almost 100% sure that this issue has something to do with my current layout..that oscillate! ..I am making a new one now and post it when I am done..

Thanks

Thomas