I am repairing a Bugera 6262 120W guitar amp which has 4 x 6L6 output tubes in a fairly typical layout. The fault is low o/p power and I strongly suspect the output transformer as DC resistance checks of the primary show a significant imbalance between the centre tap of the primary and either anode connection. One side reads about 39 ohms and the other 27 ohms. Before I lash out and buy a replacement has anyone else had a similar experience with a faulty transformer and how much of an imbalance in the resistances is deemed acceptable. All other DC voltages are well within range and max power o/p in it's current state is around 15W measured across an 8 ohm dummy load.
Such imbalances are fairly common. There are two ways to wind a center tapped winding like your primary. One is to wind half the turns, conect the tap, then finish winding the other half. This results in the same number of turns on each half, but the later ones are wound around the earlier ones, so each wind is longer wire than the inner ones. The transformer works on turns ratios, not resistance, so it doesn;t matter so much. But the longer wire means more resistance. AFter all, the resistance you measure in any transformer is just the resistance of the wire.
The other way to do it is to wind it with two pieces of wire at the same time. That way both windings will travel the same path and be of the same length. YOu then tie them together to make one center tapped winding. That results in even resistance, but is more expensive to manufacture.
If I recall correctly the Peavey Classic 30 OT measures something like 90 ohms and 120 ohms.
It is conceivable you do have a bad transformer, but it would more likely be two turns shorted together, and your meter would never detect that, so the mismatched resistance is probably not the clue. But We cannot say never...
If it were mine, I'd be checking the signal level at the grids of those 6L6s to see if they are being driven. I;'d be checking their screen grid to make sure all have B+ on them. If a tube fails, it often takes out the screen resistor for that tube. I'd check to make sure the phase inverter has B+ on both plates.
IN fact before any of that, I would plug my guitar - or other test signal of your chosing - into the FX return jack and see how that sounds. That feeds into the power amp bypassing the preamp. Strrong and clear or still weak?
FLip the speaker impedance selector switch back and forth a few times to scrub its contacts, any help?
PLug guitar into the front, and run a spare cord from FX send to FX return. ANy help?
Last, plug guitar into the front and run a cord from the FX send to some other amp and speaker. Sound OK over there or same problem? That tests the preamp.
Go over to geofex.com, and under Tech Tips find the Transformer Short Tester. It is EXTREMELY simple to make and use, costs very little for the few parts, and folloow the directions.
But I think you will find the circuit has a problem. I could be wrong. MY primary suspect would be the little transistor T1. It is just to the right of the CLEAN VOLUME control VR11A on the schematic. I have no idea where on the board it is. There are not many transistors. If it fails or is stuck on at its gate, you will have your very weak signal.
That is a mute transistor, and it quiets the amp briefly as you switch channels to prevent loud noises during transition. I'd probably just remiove it and see if the amp wakes up.
The other way to do it is to wind it with two pieces of wire at the same time. That way both windings will travel the same path and be of the same length. YOu then tie them together to make one center tapped winding. That results in even resistance, but is more expensive to manufacture.
If I recall correctly the Peavey Classic 30 OT measures something like 90 ohms and 120 ohms.
It is conceivable you do have a bad transformer, but it would more likely be two turns shorted together, and your meter would never detect that, so the mismatched resistance is probably not the clue. But We cannot say never...
If it were mine, I'd be checking the signal level at the grids of those 6L6s to see if they are being driven. I;'d be checking their screen grid to make sure all have B+ on them. If a tube fails, it often takes out the screen resistor for that tube. I'd check to make sure the phase inverter has B+ on both plates.
IN fact before any of that, I would plug my guitar - or other test signal of your chosing - into the FX return jack and see how that sounds. That feeds into the power amp bypassing the preamp. Strrong and clear or still weak?
FLip the speaker impedance selector switch back and forth a few times to scrub its contacts, any help?
PLug guitar into the front, and run a spare cord from FX send to FX return. ANy help?
Last, plug guitar into the front and run a cord from the FX send to some other amp and speaker. Sound OK over there or same problem? That tests the preamp.
Go over to geofex.com, and under Tech Tips find the Transformer Short Tester. It is EXTREMELY simple to make and use, costs very little for the few parts, and folloow the directions.
But I think you will find the circuit has a problem. I could be wrong. MY primary suspect would be the little transistor T1. It is just to the right of the CLEAN VOLUME control VR11A on the schematic. I have no idea where on the board it is. There are not many transistors. If it fails or is stuck on at its gate, you will have your very weak signal.
That is a mute transistor, and it quiets the amp briefly as you switch channels to prevent loud noises during transition. I'd probably just remiove it and see if the amp wakes up.
Thank you so much Enzo. Your reasoning behind the resistance imbalance makes perfect sense. I had never stopped to think about how the transformer is constructed but indeed the turns ratio is paramount not the length of copper to achieve it. I will do some more checking but so far all anode,cathode and screen voltages are present and correct. I ran the amp into a 100w 8 ohm dummy load with a signal gen attached and could only manage about 28 Watts. The 6L6's were glowing purple at this stage which may suggest a lot of energy being dissipated somewhere (o/p transformer?). More checks to do but I think I'll add a shorted turns tester to my test gear. Many thanks for the info and suggestions.
Problem solved! Found no volts on one of the anodes of the phase splitter 12AX7 due to an open circuit 100K resistor. A bit fiddly to replace as the whole preamp board has to be removed but once done the amp is working normally with loads of o/p and no hum. The hum should have been a clue as it was not being cancelled out as the o/p was effectively single ended not push pull. Once again thanks Enzo for stopping me from barking up the wrong tree.
I hadn't heard of this gizmo until Enzo mentioned it in this thread. Google turned up a short Premier Guitar feature (written by RG Keen himself) in which he describes it in a little more detail: The Super-Secret Transformer Tester
-Gnobuddy
The 12 ohm difference loses importance once you remember typical OT there will be some 1900 ohm Zaa (plate to plate impedance); one tube pair will see 475+39 ohm; the other 475+27 ohm.
A tiny 2.4% which of course is completely inaudible.
Overheated plates glow from dull red (sometimesyou need to turn room lights OFF to see them) to orange.
Purple often means close-to-melting screen wires, and is very dangerous ; plates are large and have relatively high metallic mass; screen wires are almost hair thin.
Unless what you call purple is actually blue glow, because of some leftover gas inside the envelope. Not dangerous.
Bifilar winding is "good" in principle because both halves match perfectly and is required in certain uses such as driver transformers but in general is avoided in OTs because the loooonnnngggg and perfectly parallel side by side wires mean a high parasitic capacitance between both halves, don´t be surprised at very high .1uF or so, and that is a highs killer or at least complicates NFB because of unwanted phase shifts at high frequencies.
Good OTs intercalate winding sections to avoid that and as a side effect may let you choose mix and match sections to get closer DC resistance ... but the more intercalation the more expensive the transformer.
Not an issue - what matters is that the number of turns is the same. The inner part of the bobbin will have less resistance as the turns are shorter. Turns count can be checked by injecting a small signal into secondary with an audio oscillator and checking the ac volts for each primary winding are the same.
The stray C is a small issue in low-Z guitar windings. The big deal is that the "loooonnnngggg ...side by side wires" have 1000+V across two layers of varnish and will eventually break-down from voltage stress. McIntosh manages to get "a lifetime" from their many-filar windings but you PAY for the extra care and fancy varnish. That's money musicians don't have for bandwidth they don't need.
A few days ago, I was testing out a doodad that does automatic frequency response measurements. I looked around for something to test, and grabbed an old Hammond 115V : 6.3V mains transformer. A small one, only 1 amp at 6.3V, i.e. rated at 6.3 VA.
For a first test, I didn't load either primary or secondary. The 115 primary was connected to the doodad's signal source (150 ohm source impedance according to the spec-sheet). The secondary was open-circuit, except for the 'scope probe measuring the voltage.
To my considerable surprise, the transformer response was basically flat to well above 100 kHz. There was a little peak of a few dB at about 130 kHz, above which the response flattened out again until my doodad ran into its maximum frequency of operation (200 kHz, I think, I don't recall exactly).
This was a transformer intended to operate on 60 Hz. Not even designed for guitar audio, never mind Hi-Fi.
I've read that small transformers tend to have wide bandwidth. And this was fed from a low-Z source. Even so, I am amazed at how much bandwidth there was.
If anyone cares, this was the doodad in question: CircuitGear Mini – Syscomp Electronic Design
IMO quite a bargain at the price, for guitar work, where the 200 kHz frequency response is more than ample. 11 bits vertical resolution vs 7 bits for the typical $400 budget digital 'scope, so traces are less jaggy. Does FFT/spectrum analysis, and the automated frequency response measurements look very handy.
The software isn't pretty and frequency response measurements don't autoscale the way I want them to, but measurement data can be exported as a text file which you can plot at your leisure on software of your choice.
-Gnobuddy
This was a power transformer - 115V AC, 60 Hz, primary, and 6.3V AC secondary. Intended purely for 60 Hz operation!
The primary was presumably intended to be fed from negligible source impedance (the 115V AC line.) The heaviest rated secondary load would be 6.3 ohms, so I suppose I could run a frequency response with a 5.6 ohm resistor wired across the 3.3V outputs just to see what, if anything, changes.
Too bad the primary doesn't have a centre-tap, otherwise I'd try it out as an output transformer in a small valve guitar amp!
-Gnobuddy
Last edited:
I have dissected plenty of transformers over the years. It is not uncommon to find two different wire diameters for each primary half. The outer half is wound with a slightly thicker wire to help equalize the DC resistances, but there is still a significant mismatch.
Years ago I bought a couple hundred Schumaker guitar amp OPT's that were made for ADA right before the fire which killed ADA. They wound up surplus, and I got my big lot for under $20 each. They are "6600 ohms to 0-4-8-16, 80VA", and speced for a frequency response of 80 Hz to 5 KHz. They have one half primary wound on the bobbin, followed by the entire 3 section secondary, followed by the other half primary in slightly thicker wire.
Measured frequency response depends greatly on the driving and loading impedances. Feed them from a pair of pentode wired EL84's and they make a decent guitar amp, but a disappointing HiFi amp. You have to pile on too much feedback to get to 20 KHz.
Feed them from a pair of 300B's and they sound quite nice and measure 29 Hz to 26 KHz at 20 watts (3db points).
Feed the same OPT's with a big fat sweep tube that has some local feedback wrapped around it and you can get 22 Hz to 30 KHz at the same 20 watts. They will handle 125 watts from 100 Hz to about 12 KHz before saturation at the low end, and high capacitance at the high end drives the tube current into the red zone. It makes for a "Metallica in the living room" experience (except for the pyro) when cranked up full into big 96 dB speakers.
Years ago I bought a couple hundred Schumaker guitar amp OPT's that were made for ADA right before the fire which killed ADA. They wound up surplus, and I got my big lot for under $20 each. They are "6600 ohms to 0-4-8-16, 80VA", and speced for a frequency response of 80 Hz to 5 KHz. They have one half primary wound on the bobbin, followed by the entire 3 section secondary, followed by the other half primary in slightly thicker wire.
Measured frequency response depends greatly on the driving and loading impedances. Feed them from a pair of pentode wired EL84's and they make a decent guitar amp, but a disappointing HiFi amp. You have to pile on too much feedback to get to 20 KHz.
Feed them from a pair of 300B's and they sound quite nice and measure 29 Hz to 26 KHz at 20 watts (3db points).
Feed the same OPT's with a big fat sweep tube that has some local feedback wrapped around it and you can get 22 Hz to 30 KHz at the same 20 watts. They will handle 125 watts from 100 Hz to about 12 KHz before saturation at the low end, and high capacitance at the high end drives the tube current into the red zone. It makes for a "Metallica in the living room" experience (except for the pyro) when cranked up full into big 96 dB speakers.
Fair enough. I can load the 3.3V side with 8 ohms, or whatever resistance I can find nearest to that (maybe 8.2 ohms).
As for the primary, what would you suggest as a suitable driving impedance? Most of the little pentodes and beam tetrodes that might be a match to this little 6.3 VA transformer seem to have anode resistances in the range of several hundred kilo-ohms.
Maybe I should just drive it with a single-transistor MOSFET or BJT current source. That should reveal the effective primary inductance at the bass cut-off frequency. I can keep DC current through the primary down to a mA or so to minimize core saturation issues.
-Gnobuddy
It's an ill wind that blows nobody any good, as they used to say in the days when people still read books.
-Gnobuddy
I don't think he actually proposed that? Left the thought hanging, didn't say it.
Yes, low-Z drive and hi-Z load makes many transformer flaws "small". The primary here is just enough Z to show some effect from winding C, but not at hundred-Ohm source. It it likely a one-bobbin winding; 2-bobbin jobs tend to show real loss above 1K (crappy coupling).
6V 6VA 1A leads to 6 Ohms secondary. 120V:6 is 20:1 of V so 400:1 of Z giving 2,400 Ohms nominal on the primary.
Try it between a 2.2k and a 6r resistor. That's a nominal reference plot. Ignoring the ill-effects of DC in the AC transformer: Using triode power tube bass will be a bit better, pentode a bit worse. Bass a bit worse taken as 3.2k:8. Low-level treble may be better with pentode and higher load Z.
I didn't propose it, because, frequency response aside, at first glance, I thought other considerations severely limited this particular transformer's use as an OT. Primary Z is too low for most small pentodes, there's no primary centre-tap so push-pull operation is out, there's no air-gap and a small VA rating so there's limited DC current capability so not much good for SE...
I believe the transformer is labelled 115V/6.3V, which makes the nominal turns ratio 18.25, and Zpri ~ 2700 ohms, pretty much as PRR says. I also think the 6.3V winding has a centre-tap; using that and one end, a more useful Zpri of ~10600 ohms should be available.
However: the secondary is only able to support 1 amp at 6.6 V. Assuming magnetic flux density in the core is roughly proportional to current and total number of turns, it should be able to support a magnetic flux equivalent to 1 amp through the 6.6V winding, which should be equivalent to (1/18.25) amps through a coil with 18.25 times as many turns...
that suggests the primary might tolerate 55 mA RMS ( 77 mA peak) of current without saturating.
That in turn suggests that the primary might support a single-ended valve biased to half of 77 mA, i.e, around 38 mA. (Current swinging between 0 mA and 77 mA at full drive.)
If that reasoning is correct, maybe this might actually be usable as a low-wattage SE OT. If I use the centre-tap and one end of the secondary 6.3 V secondary, and the primary can tolerate 35 mA DC quiescent current, and Zpri is ~10600 ohms, that translates to several watts and several hundred volts of B+.
Tomorrow is going to be a long and busy day, but I might find time to fiddle with the transformer on Wednesday. The actual frequency response measurement only takes seconds with the Circuitgear Mini, but its internal signal generator is probably limited to < 5 volts peak-to-peak, so if I put 2.7k or worse, 12k in series, there will be so little drive current into the transformer that the output signal will probably be buried in noise.
So I may have to add a single-transistor drive circuit to get the right driving impedance and still have a reasonable signal strength.
-Gnobuddy
> Assuming magnetic flux density in the core is roughly proportional to current and total number of turns,
No. Voltage.
At 115V a commodity power transformer will be high THD. You want to aim audio under 1/3rd rated voltage. So 40Vrms, 56V peak, which suggests a B+ far under 100V.
Of course depending on your lowest note and tolerance for distortion. It may be a fine thing for a 50C5 at 110VDC on *guitar*. Not working to 60Hz, not expecting "clean".
No. Voltage.
At 115V a commodity power transformer will be high THD. You want to aim audio under 1/3rd rated voltage. So 40Vrms, 56V peak, which suggests a B+ far under 100V.
Of course depending on your lowest note and tolerance for distortion. It may be a fine thing for a 50C5 at 110VDC on *guitar*. Not working to 60Hz, not expecting "clean".