Clean sound from a Peavey Windsor?

[jim p]

Tube bias voltage

Not sure which way you want to go I would think you want to take the bias voltage more negative. One thing you need to worry about is the voltage rating of the two filter caps C1 and C4 at 63 volts. If you reduce the value of R20 this will drop the voltage lower on the bias adjust circuit, I posted previously that R20 looks like four times its value in series with the pot and R23. The voltage at the diode is 77 VAC or 109 peak so a slight reduction of the value R20 will take your bias supply more negative.
Not sure how you are biasing the amplifier if you are using a scope and a waveform generator then you can look for crossover distortion. If you are just going by current and voltage on the tube then 70% to 80% of maximum plate dissipation is where you are supposed to be.
Check this page http://www.carvinmuseum.com/pdf/Ka-Boom_v1a.pdf
And this one
http://www.tone-lizard.com/Biasing.htm

Just to add to this I just got done re-tubeing a Carvin amp with four 6L6WXT Sovetek tubes and the total supply current is 150mA. I went by crossover distortion using a scope and waveform generator with the input signal at the diff amp splitter. So plate dissipation on this amp is only about 50% at idle.

On the KT88 one thing I see on the data sheets is it calls for a control grid resistor or 100k at what looks to be when plate and screen grid power dissipation is above 35 watts. That would mean that R10 and R9 in the amplifier would need to be 50k ohms which is a heavy load for the phase splitter. Even if you stay below 35 watts looks like R10 and R9 should be 100k instead of 220k to prevent thermal runaway of the tube.

 

[StringsOfJoy]

Let Me See If I Follow...

...and thank you for responding to my question...this is the only place I've seen online that has mentioned bias mods to the Windsor in serious detail, so I appreciate your sharing what you know.

I'm not using a scope and not using the crossover distortion method to bias. I don't have a scope or signal generator, and based on my reading so far I feel more comfortable using the current and voltage method using a multimeter and a Weber Bias Rite adapter: I care less about eliminating crossover distortion than I do about keeping the tubes within the relatively "safe" arbitrary zone under 70% idle plate dissipation. In fact, I want the option of going well into the 30% range (well into the crossover distortion zone and roughly where Peavey sets its stock bias) to preserve tube life even though I generally prefer biasing in the 60-70% range.

I was lucky enough to find a quad of Penta KT88s that allow me to bias just under 70% idle plate dissipation, but I have another pair that I cannot plug into the amp without exceeding the 70% "magic number". Certainly biasing into the 50% or 35% range would be out of the question with those tubes or the ones I have in right now. So that's the main interest that spurred my question.

The problem is that I'm not sufficiently experienced at understanding the circuit and the math behind it to understand which resistor values in it would have to change in order to allow me to bias ANY 4 well-running KT88, say, anywhere between 35% to 70% idle plate dissipation. I do suspect, however, that there may be more than one way to skin this particular cat.

When I was first trying to bias my amp close to 70% idle plate dissipation, I was using the amp's original Ruby EL34 BSTRs. It was impossible to get much higher than 55% with VR9 swept to offer minimal resistance.

Using alligator clips, I connected a 250k sweepable Honeywell pot in parallel with R20, at first, then with R23. I began with the maximum resistance set on the pot, close to 250k, which I thought would have the least impact on the original resistance at the corresponding resistor.

As I swept the pot toward its lower resistance settings, I noticed that effectively lowering the resistance on R20 had a minimal effect and it was difficult to find a precise relationship between sweeping the pot from max to lower resistances.

On the other hand, connecting the pot in parallel with R23 showed a direct and significant effect on the idle plate dissipation I was measuring at the tubes when I swept the pot. The negative voltage increased as I lowered the resistance of R23 by sweeping the 250k pot toward it's lower resistance settings (the tubes ran hotter), and it was clear that a parallel connection with R23 would easily allow me to run my EL34 BSTRs at a desirable 60-70% range.

It's really mainly because of this little experiment and measurements that I thought to ask first about changing the resistance value at R23 instead of R20. I reasoned that, with the KT88s in there, if I INCREASED resistance (which I couldn't do...as far as I know...by connecting anything in parallel with R23) by connecting another resistor (maybe a variable one in the form of a pot) in SERIES with R23, that would increase the resistance offered by VR9 and R23 (and the new resistor/pot) and thereby decrease the negative voltage and current flowing into the tubes. Is this correct?

If it is, what would be the math I would need to use to determine the value of the additional resistor to place in series with R23 (and VR9)? It seems like I could replace R23 with a pot or a pot and a fixed value resistor (of lower than 39k resistance) and be able to run either KT88s or EL34s in the amp at the desired range.

If I'm not correct and R20 is really the better focus for the mod, I'd like to understand why it works and how the voltage ratings of C1 and C4 being limited to 63 V fits into the equation.

Also on a side note, please let me apologize for a a confusing mistake I made above. Instead of saying that 220k for R10 and R9 would be required to use KT88s optimally (ie., squeezing the max power dissipation from them), what I meant to say was that the spec sheets seemed to call for 100k when one is running them in a circuit where KT88s will be made to dissipate more than 35 Watts max.

That's my understanding of what the spec sheets mean when they call for 100k resistors there (at R9 and R10 in the Windsor circuit). In fact, I thought I understood that it was the fact of lowering the resistance from 220k to 100k at R9 and R10 that would "open up" the tubes to potentially run at the ~42 Watt per tube max plate dissipation (and that further assumes you have the trannies to provide enough juice to the tubes). Am I right about this?

Thank you for your indulgence. As you can tell, I'm not very experienced at understanding the mathematics or design of bias circuits and specifically the one used in the Windsor. That may explain a bit if my questions seem confused or not quite correctly put. Still, I am trying to learn and try something new that I think is helping me achieve my own happy tone through this amp. I'm hoping that learning about this bit can help me understand amps better and better over time without being afraid of trying and sharing the little bits I learn.

Thank you again. I'd appreciate any bit of wisdom you can impart related to the stuff I'm asking about and I am really grateful for your response posted above!

 

[jim p]

Tube bias and control grid resistance

Just want to make sure we are on the same page relative to the cathode bias voltage on the output tubes. If you want to reduce the idle plate dissipation of an output tube you would take the bias voltage more negative say if it was -40 volts to reduce the power dissipation you would go to -45 volts as an example.
On the stock bias circuit in the Windsor the AC secondary voltage is 77 VAC equal to -109 volts peak because the filter capacitors are after R20 this reduces the DC voltage on the filter caps C1 and C4. A quick approximation of the voltage on C1 and C4 by calculation is to multiply the value of R20 by 4 then R23 and VR9 are there actual value and calculate the voltage as a voltage divider with -109 volts applied. So in the stock circuit you have 60k ohms (4 X 15k) in series with 10k plus 39k ohms total of 109k ohms. So the voltage at the top of the pot is 10k + 39K / 109k times -109 volts for -49 volts approx as the most negative voltage you can set the pot too. If you change the bias voltage circuit the most negative voltage with the stock capacitors you could have would be approx -58 with out possible shorting of the filter capacitors (they may over heat and explode if voltage rating is exceeded).

On the subject of grid resistors on tubes with power tubes you will see that there are usually two values one for fixed bias and one for cathode bias. What happens in the tube is that sometimes instead of just electrons flowing an electron with sufficient velocity hits a molecule of stray gas in the tube or impacts the plate with enough velocity to produce a positive ion. The positive ion being a positive charged particle will head for the most negative element in the tube this can be either the suppressor grid in a pentode or the control grid. When it is the control grid it will make the control grid more positive relative to the voltage drop produced across the grid resistor and increase the plate current in the tube. If too high a value of grid resistor is used the tube can go into thermal runaway where with higher plate current more ions are produced increasing plate current increasing ions, increasing plate current and cherry red plate in the tube with the possible demise of the output transformer and/or tube. When cathode biased the cathode resistor voltage will increase with plate current and give negative feedback to help prevent this thus the higher value of grid resistor that can be used with cathode bias versus fixed bias.

Why the KT88 needs such low values of grid resistance when driven hard may be due to ions produced from the velocity of the electrons impacting the plate. Also the KT88 is a power beam tube which is a tetrode with beam forming plates the same as a 6V6 or a 6L6 so there is no suppressor grid that might reduce the unwanted control grid current.

Relative to the values of R10 and R9 due to the fact they are connected to two tubes the data sheet value of maximum grid resistance will need to be divided by two. So if the data sheet called for 500k ohms you would need to use a maximum value of 250k ohms.

The value of control grid resistor used has no effect on the power capability of the tube it is attached to. Lower values will be more of a load on the preceding stage that is driving the tube. Also lower values will effect bandwidth and reduce possible blocking distortion.

 

[jim p]

Adding reverb to the Peavey Windsor head

I went with a solid state reverb drive and recovery amplifier set up for adding reverb to the amplifier I have. The reverb tank I used is the 8BB2A1B that I bought from Antique Radio Supply same as when I added reverb to the Crate V5 here http://www.thefret.net/showthread.php?t=9379&page=11
It would be best to go through the mod on the V5 posted above to see what to do relative to building and testing the reverb drive and recovery circuits. I have made a change to the recovery amplifier it looks like the best op amp to use at a reasonable cost is the NE5532. Also to reduce noise I am using both op amps in the dual package to make the recovery amplifier a differential amplifier to help lower noise pickup from the tank to the recovery amplifier. So another change that needs to be made will be to isolate the RCA jack on the output of the reverb tank as well as the input of the reverb tank. Also a ground wire will need to be added from the reverb tank to the chassis of the amplifier. Along with this if you are into making up your own cables would be to use a twisted pair inside of a shielded cable for the reverb tank output. The twisted pair should be connected to the inner and outer connectors of the RCA jack on the reverb tank while the outer shield would be connected to ground near the recovery amplifier the shield should be left floating on the reverb tank side.

To supply the op amps I made modifications to the bias circuit in the amplifier. So before you start you should measure the bias voltage on the output tubes to know where to set it to after the circuit modifications. R20 was changed to a 500 ohm 3 watt resistor and a series string of three 18 volt 1 watt zener diodes added to create a -18 volt supply for the negative rail of the op amps. For the positive supply the stock circuit has two 1N4744 15 volt zener diodes so you have +15 volts across one of them in the amplifier for the positive rail.

To sum the reverb signal into the output I added 220k ohms in series with the control grid of V3. Then the output of the recovery amplifier is connected to a 100nf capacitor and a 10k pot with the wiper of the pot attached to a 330k ohm resistor that is connected to the control grid of V3. To prevent feedback oscillation of the reverb the value of C35 was increased to 200pf from the stock value of 100pf. I located the reverb level pot on the front panel using the resonance control (bypassed resonance function in the amp). I replaced the stock pot with a 10 k pot that I got from Peavey (thanks).

As far as reverb I can get Dick Dale levels of reverb at maximum as it is set up right now. Down side there is some noise in the amplifier that I think is due to V3 having AC filament voltage (pull V3 and the amp is quiet pull V2 still have the noise) so it looks like I need to cut and hack the PCB and run wires from the preamp board to the output for the heater supply to V3.

Word of caution the op amp negative supply is part of the bias circuit supply so if it shorts the bias supply will go positive and the output tubes will be turned on hard. I scoped and measured the bias voltage on my amp with the reverb running and saw no effect on the bias voltage. However if you cause a short by backwards polar capacitor short to chassis ect. bad things could happen.

Well found where my noise was coming from it is due to ripple on the bias voltage supply. Maybe due to the added load of the reverb amplifiers and the second divider in my amp for the 6L6 tube set. I added 1100uf across the filter caps C1 and C4 with the reverb turned down now just a slight hum with your ear to the speaker. The value of caps I used may be overkill but it is what I had on hand.

If you have ever built a jfet buffer for your guitar output where you install a jfet right on the output jack of the guitar I just realized how to apply this to the reverb tank output. I have simulated a differential jfet buffer located in the tank that should balance and lower the output impedance of the reverb tank. With this as the input to a differential amplifier the noise from the tank should be close to none. I will test this out and if it works out will post the schematic.

 

[StringsOfJoy]

Thank You

It took me a while to digest what you were saying, but the general idea seems to be that regardless of how the additional resistance is accomplished it will increase the voltage on the stock capacitors (which is, in fact, what I'm trying to do elsewhere in the circuit to pump that additional negative bias voltage into the tubes.

Furthermore, the 63 V limit on the stock caps puts a lid on how much additional resistance I can create in the circuit before I risk those caps failing on me. So to get the kind of low idle plate dissipation out of my KT88s and run them cold I'd have to swap out the stock caps for something beefier (perhaps 100V caps?).

I'd forgotten to mention that I was conservatively treating the Penta KT88s as rated for 35 Watts instead of 42 Watts, so the idle plate dissipation percentages I cited above was based on 35 Watts.

 

[jim p]

A better reverb recovery amplifier

I have been bothered by the fact one of the last mods I do to an amp is to add reverb and when I do I get the 60 Hz noise back that I have been beating down. So looking into where this is coming from figure it is pickup on the reverb tank output. One of the bad things is that the output signal from the tank should connected to the recovery amplifier with a twisted pair inside of a shielded cable with the shield only tied to the recovery amplifier side. But even then the output of the reverb tank is a 1mV to 5mV signal at maximum and relatively high impedance. In all the tank installations I have isolated the input leaving the ground of the reverb tank through the reverb output shield. So a quick fix for that is to isolate the output RCA connector and run a ground wire to the chassis of the reverb tank. You still have a high impedance output from the tank so I thought of the FET buffer I have made for the guitar pickups but do to the reverb output coil floating you can even do better then that.

The buffer for the reverb tank produces a differential signal with a possible gain of up to 20 depending on the JFET used for the buffer. What this requires is to install three resistors and one to two capacitors in the reverb tank or right at the RCA jack. The remaining components are part of the recovery amplifier circuit. The circuit I built used a 2N6550 in the reverb tank but they are sold by only a few distributors and darn expensive. What you want is a JFET with low noise and high transconductance which is equal to low source impedance for the 2N6550 this is 25 to 40 mmho for 40 to 25 ohms source impedance. Instead of the 2N6550 in order of best to OK is 2SK170, BF862, BF861B and J309 the two BF fets only come in SOT23 package. What is unique about the circuit due to the fact you float the coil across the source to gate connection of the JFET you can get gain at both the source and drain side of the JFET. About the parts at the reverb tank the capacitor across the reverb output coil will cause frequency peaking so if you like twangy 2nF will do that I have no capacitor across the coil I did try from 100pF to 3nf something you can play with. R13 is to reduce the gain variation from JFET to JFET for maximum gain you can make it zero ohms. R1 is to set the current at between 1mA to 2mA if the current is high increase this if low lower it. R4 the 100k resistor could be increased to possibly increase the bandwidth of the reverb output or lowered to load down the tank for a darker sound say 50k to 25k ohms. So the source side of the JFET is connected to the outer connector of the RCA jack this is the low impedance side of the JFET while the drain is connected to the center conductor.

The recovery amplifier is a differential amplifier the gain of 10 from the JFET would be doubled if this amplifier was just unity gain but the amp has a gain of 10 so total gain is 200. So the circuit to the JFET is completed by the parts at the recovery amplifier R3, R2 R12 and R11. C1 and C2 along with R9 and R10 provide a high pass filter to help keep the low frequencies out of the recovery signal. C6 across the input will limit the high frequency input to prevent hearing your local AM radio station (because some of the JFETs are RF FETs). The first op amp U1 is to provide equal gain and high input impedance for the negative input to U2. The op amp can be an NE5532 dual op amp it has good low noise numbers and will work well for this application.

To sum it up I have a little hum but the power supplies in the Peavey Windsor I added this too are not great on the low ripple side I should add regulators.

Remember if you have both the input and output of the reverb tank isolate you will need to connect a ground wire to the reverb chassis. Also if you have a painted tank like a Ruby need to break the paint where the four springs that suspend the reverb assembly are attached to complete the grounding.

I want to correct one thing above the Jfet buffer is acting as a current source at both the source and drain so the impedance at the source and drain is the same 1 k ohms in the circuit as shown.

 

[svenbyliner]

Mod Idea

Hi Jim, I found this thread looking for some mod ideas for my windsor head and was not disappointed! A very thorough and interesting read.

Anyway, I recently had an idea for a relatively simple mod for my amp, which is an almost stock peavey windsor head (I changed out the 12AX7 preamp tubes to lower gain tubes.)

I've attached a picture to this post of what I'd like to do it. Simply put, I'd like to create a second preamp volume knob for the amp. This way I could set the two preamp volume knobs to different settings and switch remotely between the two.

Anyway, I'm a novice at this stuff (I've done mods but never designed one) so any help or advice is greatly appreciated.

In particular I'm wondering If I should put the second preamp volume knob on the amp itself, or if it's possible to put it on the pedal (to be used as the switch.) Also, how to best connect the pedal switch and the amp.

 

[jim p]

clean and overdrive?

Just having one added pot in the preamp section would not be the same as having a clean and overdrive (dirty) channel if that is what you are trying to do. It would be similar to the boost that is in the amp with an adjustable level. You could do almost the same thing as adding one pot by just using a volume peddle between the amp and the guitar. If you cut the boost relay free and added two pots you could have two preamp gain settings and two volume settings to make the amp more like a two channel amp. The resonance pot is 1 Meg I think and I find the control to do little for the tone of the amplifier so it would work well as a second volume (level) control for the overdrive. Then you would need to add another pot in parallel with the preamp volume pot for the overdrive level. But one problem that may occur with this mod is popping on the speaker when switching between channels. The relay contacts may bounce or break before make and cause this to happen.

Another thing you can do with out mods to the amplifier is to use the loop in and out as a switchable volume control using the footswitch along with an input level control on the signal into the amplifier. So you turn the loop on and send a high level signal into the amplifier that you can control the volume on with a volume pot added to the effects loop same as the level pot with overdrive. When you turn the loop in/out off you input a low level signal into the amplifier for your clean sound that is controlled by the master volume on the amp.

 

[svenbyliner]

cont.

Yeah, the basic idea is to simulate a clean and overdrive channel.
My idea to control the volume of the "OD" channel was to simply wire a second pot onto the pedal I'm using (which has two good jacks and a good switch but a busted board, which is why I want to salvage it.)

Now, the second pot would be wired like a normal volume pedal for a guitar and would be placed into the signal path in the effects loop (thus "acting" like a master volume by coming after the preamp stage. And yes, this means the FX loop has to stay on all the time, but I've got an ab box that can go after the DIY switch pedal and act as an on/off for the rest of the FX loop)

This seems like the simplest way to control both the gain and volume from one pedal, though it will of course require a third output jack for the pedal (one jack to connect to the preamp, and one in and one out to connect to the effects loop signal chain.)

That part of the equation should be simple, I'm just unsure about how to wire the preamp gain pot so that it will be switchable.

I haven't had a chance to take the back panel off and look at it yet, but does the volume pot simply have two wires connecting to it that I need to splice a switch between?

If that's the case then I can easily hook up the pot and the two wires in question to a 4 conductor TRRS jack (with a matching jack installed on the pedal) then I can place another switch in the line before the jack, which would allow me to turn the new feature on and off from the faceplate of the amp.

I'm guessing though that the foot pedal I use will need a new switch to be able to work both the preamp volume and master volume simultaneously.
What would I use, a 4pdt?

 

[svenbyliner]

also

Do you think a noise gate will have any effect on the possible "popping" noise this mod might cause when switching between preamp controls?

If not, any thoughts on a buffer circuit?

 

[Muziekschuur]

Windsor Studio on the way....

Hi,

I have a Windsor Studio on the way. I allready have a ECC81 ready to put into V1 and maybe one asswell in V2. If that won't cut it I allways have a V-twin (Mesa Boogie) available to do some MIDBOOST cutting.

I allso have two 1x12 Trace Elliot openback cabs with Celestion superleads in them (16 ohm). So let's see what this will bring to the table.

I allso have two Fender M80 combo's wich I will use together with the Peavey for a WET DRY WET setup.

We'll see what that brings to the table.

Cheers,


Muziekschuur
The Netherlands

 

[BigBadWolf1171]

i own one of these windsor heads, the clean isnt the best, i personally like it. i run 3 -9 volt batts on my emgs and with the help of a 15 band EQ i can get decent cleans from it,just before it starts to break up a little.. i have no knowledge of the internals of an amp, like you fellas do,(im envious!).. i personally would love to get more gain from mine, but im clueless. id imagine i could change tubes but i really wouldnt know which to get..

 

[aurban18]

Why i can't see the pictures?