Transfomer ID Help

[honda2000]

I have a couple of output transfomers that were in a amp my dad built. They have 4, 8, 16 outputs and they were used with a push-pull 6v6 setup.
The only id on the transfomers are T582-113 089.
I can't seem to find any info on the web.

Thanks,
Chris

[honda2000]

I do know that there's only 3 wires on the primary, I believe it's 10k ohm center tap. Will these work in this project?

Chris

[erichayes]

Hi All,

If the transformers were used with 6V6s in push-pull, they're probably rated at around 8000Ω and can handle roughly 75~100 mA of current. They probably don't have screen grid taps, but if they weigh more than a pound apiece they're probably fairly decent. They could be used with 6AQ5s and 6BQ5s in PP, as well as 6V6s, with good results, and would probably work fairly well with 6BM8s or 6GW8s. Experimentation is what DIYing's all about.

[honda2000]

Thanks for the reply Eric. The schematic that the amp was built from called for a transformer of 9k to 10k, that's what I was going by since I don't have the specs on it.
No they don't have the screen taps, I would have no idea how else to connect the outputs of the tubes to them...

Chris

[Shannon Parks]

Thanks for the reply Eric. The schematic that the amp was built from called for a transformer of 9k to 10k, that's what I was going by since I don't have the specs on it.
No they don't have the screen taps, I would have no idea how else to connect the outputs of the tubes to them...

Chris

Hi Chris,

If you own horn or other high effieciency speakers, I would recommend pseudo-triode connecting them. I have been listening to one of my ST35's in triode mode the last several weekends and really like it. But ~5W doesn't get you far. Other than that, pentode connection would be the way to go. You can tap into the next filter stage after the B+ and run that to all your screens. You'll need to do some V=IR math and resistor changes to ensure you still get 300V to 320V at the last filter stage.

Shannon

[honda2000]

Shannon, thanks for the ideas!!! I assume the B+ is the center tap of the output transformer? I'm tyring to learn here...

Chris

[EWBrown]

The primary should have three leads, the usual color code is red for B+ center tap, and brown and blue for the two plates. If in doubt, ohm it out...

/ed B in NH

[honda2000]

Thanks EW and everybody!!!
The primary's have a red, blue, and blue/yellow, or could be a white trace.
When I checked the impedance, red to blue is 325 ohms, and red to blue/yellow is 400 ohms. Blue to blue/yellow is 730 ohms. They are the same for both sets of transformers. I guess it doesn't make a difference which lead goes to which tube as long as the b+ goes to the center tap?
I was doing a search around the shop, and I came up with all new caps for the original 6v6 circuit, along with a bunch of the original resistors, new tube sockets. All I'm missing a $10 worth of resistors from what I can tell.
I was thinking of building a turret board from the schematic, power supply first, then the rest of the circuit. It was originally done point to point, and that's why I ripped it apart.


Chris R.

[erichayes]

Hi All,

Chris, you measured the DC resistance of the primary, rather than the impedance. The reason for the discrepancy in resistance between the two halves of the winding is simple physics: although the impedance is based on the number of turns, the DC resistance is determined by the overall length of the wire to achieve those turns. Windings closer to the bobbin require less wire, and, hence, have lower resistance. I won't go into such esoterica as interleaving right now, but this is an easy way to determine which half of a tapped winding is the "start" and which is the "end" (very handy when working with negative feedback).

BTW, I use point-to-point wiring in all my amplifiers for the simple reason that it sounds better than printed circuitry and, to a lesser extent, turret board construction. My concession to that rule is to use PC boards where they aren't involved with the signal path (power and bias supplies, monitoring circuits etc.). I've been working with Bud Wyatt for the last two months on some amplifiers that are going to terrorize the planet--today we did a PPP KT88 monoblock that does 121 watts @ 0.5%THD before tweaking the feedback--and they're all point-to-point.

I'll relate my adventures with Bud in a later thread (more likely a rope).

[Shannon Parks]

Chris, you measured the DC resistance of the primary, rather than the impedance. The reason for the discrepancy in resistance between the two halves of the winding is simple physics: although the impedance is based on the number of turns, the DC resistance is determined by the overall length of the wire to achieve those turns. Windings closer to the bobbin require less wire, and, hence, have lower resistance. I won't go into such esoterica as interleaving right now, but this is an easy way to determine which half of a tapped winding is the "start" and which is the "end" (very handy when working with negative feedback).

Hi Eric,

This reminded me of a question that I had a few weeks ago regarding leakage current affecting the primary inductance. I was doing some ST70 tests and had the EL34s warmed up and the bias was matching right at 50mA each. I did my tests and had slighty poor 20Hz performance. I then set my audio analyzer to 20Hz and started tweaking the bias balance pot (not on stock ST70, but on my board) until the 20Hz THD was lowest. FWIW, I think it dropped from .85% to .5%THD. Anyhow, the tubes were balanced at about 51mA and 49mA at this point. So my thinking at that time was that the this was due to uneven leakages or an inbalance in the A470. Since this is DC leakage, maybe this is directly related to the DC resistance? Just guessing here. I guess I can test this out on a few other trannies and use a couple sets of EL34s.

Shannon

[honda2000]

Eric, thanks for the info, I didn't know the exact reason, but that's kind of what I thought.

Chris

[EWBrown]

If you want to get really critical about balancing the output stages, there is bias matching, and there is also "AC" balance, usually the plate resistor of one side of the phase splitter has a pot in series to allow for fine tuning, the other way is to use (like on the ST70 driver board) a 1 meg pot to balance the G1 voltages.

The usual method to set AC balance is to feed in a 1 KHz sine wave, and then set for zero volts AC between the two output tubes' cathodes. Use a dummy load, not a speaker, in order to maintain "the peace"...

On a slightly different tangent, would there be any advantage to using fixed bias rather than cathode bias on an ST35 design? I would SWAG that the B+ should be reduced accordingly, by around 15 VDC as the cathodes will be within a few tenths of a volt to ground. Since I can never leave anything "well alone" I was cooking up the plans of ST35s using a 5HA7 (mentioned earlier elsewhere) or using two 6AV6s and two 6C4s in place of the 12AX7 and 12AU7. Of course 6GK6s have already been tried and tested. Or 8CG7s and 8BQ5s. Plinkers rule! Yellow_Light_Colorz_PDT_11

/ed B in NH

[EWBrown]

If you want to get really critical about balancing the output stages, there is bias matching, and there is also "AC" balance, usually the plate resistor of one side of the phase splitter has a pot in series to allow for fine tuning, the other way is to use (like on the ST70 driver board) a 1 meg pot to balance the G1 voltages.

The usual method to set AC balance is to feed in a 1 KHz sine wave, and then set for zero volts AC between the two output tubes' cathodes. Use a dummy load, not a speaker, in order to maintain "the peace"...

On a slightly different tangent, would there be any advantage to using fixed bias rather than cathode bias on an ST35 design? I would SWAG that the B+ should be reduced accordingly, by around 15 VDC as the cathodes will be within a few tenths of a volt to ground. Since I can never leave anything "well enough" alone, I was cooking up the plans of ST35s using a 5HA7 (mentioned earlier elsewhere) or using two 6AV6s and two 6C4s in place of the 12AX7 and 12AU7. Of course 6GK6s have already been tried and tested. Or 8CG7s and 8BQ5s. Plinkers rule! Yellow_Light_Colorz_PDT_11

/ed B in NH

[leadtower]

Say E.W. Can you tell us how you would go about developing fixed bias on the ST 35? I have been discussing this on another forum and your input would be greatly appreciated.
Thanks J.R.

[erichayes]

Hi All,

I realize I'm not Ed, J.R., but since my 6BQ5 amps use fixed bias I think I'm qualified to throw my two cents' worth into the mix.

I'm only going to deal with one channel for simplicity.

First, remove R17 and R20 and replace them with wire jumpers. If you're paranoid about heat dissapation replace R43 and R44 with 10Ω 1W 1% resistors (Mouser #71-RS1A-10), or if you want industrial strength fixed bias, 1Ω (71-RS1A-1).

Lift the grounded ends of R13 and R14 from the board and connect them together with a length of insulated wire. This junction is where the negative bias voltage will be injected.

The simplest way to build a bias supply is by using a separate transformer. A Mouser 41FJ300 will work nicely, providing 18VAC @ 300mA--'way more than necessary. You'll also need a 1N or UF rectifier of at least a 50 PRV rating (if you bought extra UF4007s when you built the kit, use one of them). If you want to hear the difference between half and full-wave bias--and you will--get a small full wave bridge module (Mouser 583-RB152), a couple of 100µF @ 50WV capacitors, and a 3.3KΩ 0.5W resistor to complete the supply.

You'll also need a 50KΩ linear taper pot (31VC4050) and a couple of 15KΩ 0.5W resistors per channel.

Method of construction is up to you, but I'd suggest perfboard as probably the easiest overall technique.

If you're using a single diode, connect its cathode (banded end) to one of the secondary leads of the transformer. Connect the other secondary lead to ground.

If you're using a bridge, connect its "AC" or"~" leads to the transformer and ground the "+" lead.

Connect the remaing lead of either the diode or bridge to the negative lead of a 100µF cap and one lead of the 3.3KΩ resistor. Ground the cap's positive lead.

Connect the other lead of the resistor to the negative lead of the second 100µF cap; ground the positive lead of the cap. Connect two of the 15KΩ resistors to the junction of the 3.3K and second cap. Connect the other leads of the resistors to one of the outer lugs of the two pots.

Take the remaining 15K resistors and connect one lead to the other outer lug of each of the pots. Take the remaining leads of the resistors to ground.

Before connecting the ground of the power supply to the ground of the amp, apply line voltage to the primary of the transformer, connect a DC voltmeter to the wiper (center lug) of each pot and ground and confirm that you have a negative variable voltage with respect to ground. Just run the pot through its rotation and see what numbers you get. They should vary from roughly -20V to -5V. If you get these voltages, set each pot to whichever extreme gives the maximum negative voltage, disconnect power and connect the grounds.

Hook the bias supply transformer primary to the power switch of the amp (just hook it in parallel with the amp's power transformer). DON'T FORGET TO DO THIS! If you don't, your 6BQ5s will go into Roman Candle mode.

Connect the wiper of the pots to the R13/R14 (R15/R16) junctions created earlier, using hookup wire. Connect a DC voltmeter across R43 or R44 and adjust the corresponding pot for 250-320 mV if you stuck with the original 10Ω value, or 25-32 mV if you went with the 1Ω mod. Do the same with the other channel.

My experience with both Sovtek and JJ 6BQ5s is that they can take plate voltages well in excess of their ratings. I'm running the 1773s at 375V on the plates, -13V on the grids and 32mV on the cathodes. Out of 20 amps, I've had zero output tube failures over the last five years. Those puppies are tough.

After the dust settles, I'll talk about DC and AC, or dynamic, balance, which Ed Brown touched upon earlier...if anyone's interested.