Power Amp Build Problem

[silverfox]

For my first build I've taken the parts from a 100 watt Carvin MTS3200 and reassembled them to build a 35 watt, EL34 based version of the power amp section. No preamp or channel circuit build. This worked well before I took the Carvin apart.

The power amp circuit I built follows the schematic as drawn with the exceptions of: The transformers came out of a 35 watt 7686 based amp and since the power transformer had a 50 volt tap I put together a full wave bridge bias circuit. On the full wave bias circuit, I cut the junction where the original MTS bias circuit patched in (220K resistors) and put a bias feed on either side of them.

I got only a reasonable level of background noise and a signal placed on the input appeared, un-amplified, on the output in a reasonably undistorted state. So the signal found it's way through the circuit.

Prior to getting the bias set to a reasonable value, -37 volts DC the tubes wanted to red plate but I kept the power level down with the variac. I based the -37 volt setting on the data from the schematic for the bias circuit I used. While there may be a bias level adjustment needed, I think the problem is in the design or perhaps the output transformer application- Design problem.

All parts worked in both amps before removal. The power supply voltages are close, within 15 volts.

MTS schematic: http://www.el34world.com/charts/Schematics/files/carvin/carvin_mts3200.pdf


Current State:

After setting the bias and using the variac, when I start bringing it up, at around 60% of line on the primary- 80VAC, it begins to sound like an AM radio station that is un-tuned just screaming random noise. Nothing discernible.

I immediately shut it down since it appears it is going to get really loud and perhaps fry something. The sound peaks until the voltage begins to drop off and then fades to black.

Removing the AX7 driver tubes does nothing to stop the noise. It does go away if I remove the output tubes.

Since the noise increases in intensity I get the sense that the amp is working in a primitive sense- Perhaps?

Since this is just a first prototype build with used parts I don't know how much effort I'm going to put into it but I would like to get it going to see if the output transformer is going to work with EL34 tubes.

Any suggestions as to what I should check to resolve the problems?

Silverfox

[firemedic]

You may have the OT primary wires reversed. Try switching them.

[eleventeen]

Concur w/firemedic.

[silverfox]

New to this so- What difference does it make. To me it just looks like a coil. But now I'm thinking it has something to do with the phasing of the feedback circuit?

I'll switch the leads to check.

Thanks.

[eleventeen]

Your suspicion is correct. If incorrectly phased, the NFB connection turns the intended NEGATIVE feedback into POSITIVE feedback. Unless you are building multiple examples of an amp, using identical OTs and know that "brown goes to this side, blue goes to this side", there's no real good way to determine which way the OT gets wired in advance, so it is generally considered good practice to leave extra lead length and temp-solder the OT primary leads until correct phasing has been established. That's really all there is to it.

[silverfox]

Okay. I swapped the leads and now it works but poorly.

First of all, Until I plugged an instrument cable into the input and then plugged that into a stereo headphones out jack, there was way too much AC interference noise in the amp- lights etc.. Plugging the input into another active source cleared that up so I'll have to check the ground issue. It is not grounded to the chassis. Just the circuit ground and the jack is plastic so there are no chassis connections. There is a chassis ground but I purposely didn't connect it to avoid putting chassis noise on the input.

Another question- There is a faint blue glow in the tubes. The plates may be getting close to redplatting. Not sure. It seems as though they are getting too bright at full power now. Should there be any other red glow other than the filament glow?

I'm going to connect some 1 ohm resistors up and check the bias that way since there seems to be some sort of distortion at the higher frequencies.

Any suggestions regarding these problems?

Silverfox

[HotBluePlates]

... the power transformer had a 50 volt tap I put together a full wave bridge bias circuit. On the full wave bias circuit, I cut the junction where the original MTS bias circuit patched in (220K resistors) and put a bias feed on either side of them. ...

Tell me more about what you mean by "full wave bridge bias circuit".

Does your PT have 2 wires for the bias tap (the way the Carvin has two 8v taps for the relay power supply)? Or does it just have the bias tap and the high voltage CT?

If the latter, you have a diode in the way of your high voltage center-tap, and will probably never get decent operation. Again, if there is only 1 50v tap, re-wire the thing like every other Fender, but test power supply operation without any output tubes to see where the B+ and bias voltage land.

Another question- There is a faint blue glow in the tubes. The plates may be getting close to redplatting.

Turn the lights off and look at the plates. If they redplate, you'll know it because you'll see cherry red in the middle of the plates, far from the ends where you see the filaments glowing orange.

Forget the blue glow. People have been obsessing about that for decades, but in probably 99% of cases using modern tubes, it's just an artifact of the glass used in the envelope.

[silverfox]

Tell me more about... Well you see it was foggy that night, just after a rain but I could clearly see...

Oh you me the other bridge.

The tubes do not appear to be red platting. Just a bright orange glow from the filaments.

Okay, I used a different power and output transformer rather than the original parts from the Carvin. One from a Mcgohan 35 watt PA. There is a 40 - 50 volt winding with a center tap that was used on the mcgohan for a bias circuit. Did not connect the center tap. I just went back to look at the Mcgohan circuit and realized that circuit has the bias supply center tap grounded but it is a half wave bridge. Hmmm.

I constructed a full wave bridge bias circuit with a 2500 ohm resistor, a 25uf cap, two pots and two 10K resistors. Should be similar to sluckeys amp scrap book version. Reverse polarity wired the cap to ground, connected that ground to my circuit ground in the power supply section. I then adjusted the outputs of the bias circuit going to each of the two resistors at "E", for _37 volts. I broke the junction connection at "E" so that each bias tap feeds one of the two resistors (each tube). Maybe I need a bigger cap. Because I only had a 50 volt cap I connected the 25uf cap after the 2500 ohm resistor instead of before. A better cap is on the way and I'll rebuild the bias circuit.

Here are some voltage readings with the amp on and a signal on the input. I have also listed some minor changes to the supply design after the voltage readings:

Input voltage:                 119 VAC

1st Input Cap- Plate B+    436

Screens:                        427

TP "C"                            420

TP "D"                            413

Bias Voltage                    -37

Changes to supply: I added a Fender choke, Bandmaster size between the first two input caps. The first two input caps are 47uf 500VDC each. I did not put in the next 22uf capacitor. The remaining supply is as drawn with the last 3 22uf caps present.

The amp is picking up 60 cycle hum like crazy.

My power situation is a little different than usual in that I am off grid. I use a high quality inverter that has not caused any problems with my other amps. However, this amp has it's guts hanging and it is plugged into a variac at the moment so maybe I'm picking up noise from that. Again, this being my first build the grounds may not be done perfect. I have to check that tonight when I put cathode resistors in.

I used shielded cable for all my input and signal paths to the tubes. Perhaps over did it. That is the only connection to the chassis and then just one end of the shield.

I stress again- I did not ground my circuit to the chassis. Is that a problem?

[silverfox]

Hurray! My first build is working. I think the bias can use some adjusting. I grounded my circuit ground to chassis ground and all the interference went away.

It definitely likes the Marshall and other bass-heavy metal pre amp tones. Must be that old iron. Seems to give it a sag sound even though the rectifier is solid state.

Thanks everyone for helping me get through my first build.

One thing I think I've learned from other experiences- It's working, sounds great! DON'T START TINKERING WITH IT... not yet anyway.
Still have to build the preamp and other sections but it works great with an amp emulator.

Thanks again everyone!
Silverfox.

[HotBluePlates]

I grounded my circuit ground to chassis ground and all the interference went away.

You probably have your power cord's 3rd wire attached to the chassis, right?

The thing to do in your case is put a 0.01uF 1000v cap in series with a 51Ω resistor, from the circuit ground to the chassis. Then, place a SPST switch across the series components.

Most of the time, you will set the switch so it is a short-circuit across the series components. That's because you'll have no hum/buzz in that setting. However, if you ever plugged in other gear and created a ground-loop, you flip your switch to only connect the circuit ground to chassis ground through the series components.

See, that would kill ground-loop hum because the circuit ground is no longer connected to the power cord 3rd wire. But you amp is still finding that ground through the other part of the ground loop (which would be broken, so call it a ground "c"  ).

Glad you got your buzz issue under control.

And did you mean that the bias winding is it's own complete winding, with no connection whatsoever to the high voltage winding?

[silverfox]

For the grounding of the line circuit- The line ground goes to chassis. I left all the Carvin stuff on the little board that mounts the power cord adapter then put the ground wire to chassis.

I just had circuit ground floating above chassis ground.

The bias winding is a separate winding. I ohmed everything out when I removed the trannys. I ordered the schematic for the Mcgohan when I first bought the amp at a 2nd hand store for ten bucks. The 7686 tubes were worth more then that.

I'm planning on getting to everything tomorrow. Had to town to today.

This build used parts scrapped from 4 amps and it is a concept build. I don't think I want to try to cram everything into the chassis. I could but it just wouldn't look right. I've still got to put the preamp and other circuits together but in the interim  it works well with my Digitech RP 355. Models Marshall really well. Nice deep grinding low end distortion.

Despite capacitance issues suggested with using too much shielded cable I've found on this build, so far, all signal path work well. It seems very quiet now that the ground issue is gone. Just a background hiss. If there is any interest I could put up some audio. It's just a power amp though. Unless of course you wanted to hear something Hendrixy... But then everyone does that stuff.

I'll try the ground switch. The Trace I took apart had that. Ground lift circuit is what they used. Not sure if that is what you are saying but I'll check.

I've still got 3 1/2 tube sockets left too.

Silverfox.

[eleventeen]

"when I first bought the amp at a 2nd hand store for ten bucks."

Ahhh, for those days. Very rare now. Though I bought a Radio Craftsman amp for $25 and sold it for $685 last year. I don't think there are as many of these going into homebrew builds as there are going straight to scrap metal, sadly.

[silverfox]

Setting the bias:

As previously stated this bias circuit is powered from a separate winding of about 50 volts rectified DC. The rectification is from a bridge circuit.

The bias source circuit consists of: rectifier-range Pot -25uf 50v el capacitor- 25K bias pot. The bias adjustment is taken off the wiper of the bias pot. The far side of the bias pot is to ground through a 10K resistor.

While setting the bias, several times now I get a runaway situation where the bias of the tube I'm not setting takes off. The first time I didn't catch it right away and the tube Red Platted briefly.

It seems as though a fulcrum is moving off center point on a Teetertoter if you get what I mean. Adjusting one pot moves the fulcrum center and the other end of the Teetertotter- bias current, goes through the roof very quickly.

Don't understand this at all.

Since this began to happen I use the variac to bring everything up real slow and keep adjusting the bias in very small increments to keep anything from taking off.

I have set the bias between 25 and 31mv across a 1 ohm cathode resistor- for EL34 tubes. I want to get away from crossover distortion. The plate voltage is +436. The schematic is in the first post if you haven't seen it yet. The voltages in the amp are lower than the schematic voltages

Questions: Is this normal to have the bias current take off for the moon while it is being adjusted?

After doing some research it appears the bias should be set at closer to 40 ma for 70% power with the B+ I've stated (Tube power rating * % power desired / B+  = mv meter reading).   Correct?

I have another question:

Does the bias voltage have any bearing on the tone attribute? Perhaps the bias voltage will always be x if the ma bias setting is y. I don't know.

Once I get the bias set up properly, how likely is it to runaway under normal operating circumstances?

Silverfox.

[silverfox]

The bias can be adjusted and once set it seems to float around between a couple mv as I'm playing. It just seems unstable and hard to set. Probably the design.

I'll be out until tomorrow. I'll get a schematic up as soon as I can.

Thanks.

[silverfox]

Problem resolved.

Strange as it is, there are two bias pots. I was monitoring one tube but adjusting the other tubes bias pot. The divider circuit went over the edge causing one of the tubes to run away.

The adjustment process proceeds normally now.

The circuit is similar to the full wave bias circuit in the amp scrapbook. Only using one capacitor though.

Is there any bearing on the bias voltage to the tone attribute or is it strictly a current issue? In other words, could the bias voltage be set at -21 vdc and 42 ma per tube and have the same sound?

Silverfox.

[alerich]

Strange as it is, there are two bias pots.

I prefer the independent bias pot approach. I modded my Mig60 for dual independent bias and the JCM 800 I am building will include them as well. I'm a little confused here, though. I was looking at a schematic for the MTS 3200 and I only see one bias control. Are there different models of this amp floating around with slightly different designs?

[silverfox]

At the beginning of the topic you will find this is a rebuild of the MTS 3200 power amp section. I had one, didn't like how it sounded and scrapped it out. As a first test build, I rebuilt from schematic, the power amp section only.

I'm finding that I really don't care that much for this build and I'm glad I reused the parts from the original amp. The tone has too much bass in it and it sounds kinda fuzzy distorted. Bias currently set at 42 ma/tube.

I want a simple 50 watt good, clean sounding power amp as my bench test for preamp builds. Any suggestions?

Silverfox

[Willabe]

As a first test build, I rebuilt from schematic, the power amp section only.

The tone has too much bass in it and it sounds kinda fuzzy distorted. Bias currently set at 42 ma/tube.

I want a simple 50 watt good, clean sounding power amp as my bench test for preamp builds. Any suggestions?

Can you post a schematic of the build as it is now?


            Brad     

[silverfox]

The link is in the first post at the top. The bias circuit is working. It is the only thing changed. It is a full wave bridge circuit similar to sluckeys amp scrap book design.

If you follow the path on the MTS 3200 schematic picking up from the return signal, that is the amp as drawn. The only exception is I only wired up two power tubes and the power and output transformers are from a 35 watt amp. The voltages are off by about 10 volts I believe due to the addition of a choke to the power supply after the first input capacitor. I also used 47uf caps for the first two input caps in the drawing. Further down in the posts you will find a chart of the voltages.

http://www.el34world.com/charts/Schematics/files/carvin/carvin_mts3200.pdf

I've got so much going on right now I don't have time to learn and start using the schematic tools- jschem. My computer just crashed and I've got that to rebuild. On an alternate right now.

I'm thinking of using this schematic for a power amp build- Just the power amp section, perhaps a master volume version.

http://www.projetg5.com/Projets/MI60/MI60_EVO_beta3.pdf

Fox.