I suspected my AC volt meter has a hard time measuring single digit millivolt AC voltages. So I increased the current sense resistor to 2K, as well I obtained a true RMS current meter. I performed the tests both ways--using the current sense resistor to estimate current, and using the AC ammeter in circuit--and got only 1% difference in results across the board.
I also thought it would be good to baseline these tests against my own sample of a known good PP OPT. I have a new pair of Dynakit Parts A470 OPTs that I recently purchased for an upcoming build, so I used one of them as the baseline.
Here are the results:
Now what do you make of these results?
I also thought it would be good to baseline these tests against my own sample of a known good PP OPT. I have a new pair of Dynakit Parts A470 OPTs that I recently purchased for an upcoming build, so I used one of them as the baseline.
Here are the results:
Code:
[H="5"]V L(A470) L(Northlake)[/H]
1 63 78
3 137 134
5 185 167
10 260 230
24 366 310
48 458 388
96 558 471
130 610 510Now what do you make of these results?
It looks like a P-P OT now. Very mysterious.
Could this amp really work?
How much current would the tube have been drawing through this? A severely current starved tube maybe?
I am left with extreme scenarios.
Outer space_1:
Could the laminations be some special alloy that can be permanently magnetized to counter the DC bias? Perhaps by a power up pulse, like color TV sets used to demagnetize the picture tube/chassis.
Outer Space_2:
A HF flyback inductor (powered by a HF SS switch, pulsed) was pumping opposing "DC" current into the OT to counter the tube DC current.
Outer Space_3:
A HV Mosfet was used as an opposing current source.
Outer Space_4:
Some new configuration never seen before is hidden in there. Like a Matrix configuration with two tubes and two OTs combined in pseudo P-P to cancel the DC but pass two separate AC signals through. Sum and difference of the secondaries to retrieve signals.
Outer Space_5:
Two tubes are used per OT in a Circlotron configuration, one tube without AC signal (just acting as counter CCS)
Outer_Space_6:
The laminations are coated with a thick non-magnetic coating so as to force all the magnetics to pass through the butt joints. (feeble air gaps)
Outer_Space_7:
The I lams are fake, cardboard/fiberglass.
Could this amp really work?
How much current would the tube have been drawing through this? A severely current starved tube maybe?
I am left with extreme scenarios.
Outer space_1:
Could the laminations be some special alloy that can be permanently magnetized to counter the DC bias? Perhaps by a power up pulse, like color TV sets used to demagnetize the picture tube/chassis.
Outer Space_2:
A HF flyback inductor (powered by a HF SS switch, pulsed) was pumping opposing "DC" current into the OT to counter the tube DC current.
Outer Space_3:
A HV Mosfet was used as an opposing current source.
Outer Space_4:
Some new configuration never seen before is hidden in there. Like a Matrix configuration with two tubes and two OTs combined in pseudo P-P to cancel the DC but pass two separate AC signals through. Sum and difference of the secondaries to retrieve signals.
Outer Space_5:
Two tubes are used per OT in a Circlotron configuration, one tube without AC signal (just acting as counter CCS)
Outer_Space_6:
The laminations are coated with a thick non-magnetic coating so as to force all the magnetics to pass through the butt joints. (feeble air gaps)
Outer_Space_7:
The I lams are fake, cardboard/fiberglass.
Last edited:
Okay. Well it was an interesting journey for me to learn how to measure primary inductance. I’m going to put them aside now and move on to the next project. But now my interest is piqued about building some 845 SE amps. I need to obtain some proper output iron. I was looking at Electraprint. They make some 9k 20 watt iron that I think would work. Anyway, I appreciate the help!
#8
The I lams are permanent magnet material designed to counter the DC current.
Plan 9 from outer space:
We are being evaluated on our solving an impossible problem. Maybe this amplifier was called the "Enigma" 🙂
You might see if you can borrow a tube to see how well one channel worked.
The I lams are permanent magnet material designed to counter the DC current.
Plan 9 from outer space:
We are being evaluated on our solving an impossible problem. Maybe this amplifier was called the "Enigma" 🙂
You might see if you can borrow a tube to see how well one channel worked.
Last edited:
#9 Inductance drops substantially with DC current but plateaus to something that is still useful. I've seen a thread somewhere else where this was demonstrated, a PP transformer was being used for SE and it worked OK, even though the theory says it shouldn't.
Yeah, I can see where the primary L would just drop from DC saturation until the low residual permeability acts like the usual air gap to prevent further drop. It probably will eat up more than 50% of the available core flux unfortunately, where a proper gap is tuned to just use up 50% from expected DC current. (so the signal can swing symmetrically around the 50% level, 0 to 100%)
I think the usual SE OTs are wound with more turns than the P-P case to make up some for the low permeability. So using a normal P-P OT would likely end up with reduced LF response with the DC in it. In which case using a P-P OT with abundant primary L, like the Edcor measured above, would be helpful.
Depending on the DC saturation to set the pseudo "gap" however will allow the inductance to vary more with signal than with a real gap. Probably make for plenty of 2nd (or 3rd) harmonic from asymmetric magnetizing current (which is at 90 degrees to the signal phase, so may end up as 3rd H) .
I think the usual SE OTs are wound with more turns than the P-P case to make up some for the low permeability. So using a normal P-P OT would likely end up with reduced LF response with the DC in it. In which case using a P-P OT with abundant primary L, like the Edcor measured above, would be helpful.
Depending on the DC saturation to set the pseudo "gap" however will allow the inductance to vary more with signal than with a real gap. Probably make for plenty of 2nd (or 3rd) harmonic from asymmetric magnetizing current (which is at 90 degrees to the signal phase, so may end up as 3rd H) .
Last edited:
I was curious to see how well this scheme of SE use of a P-P OT might work, so I measured the magnetizing current of the Edcor CXPP100-4.2K at max rated power, 100 Watts.
120V on the 16 Ohm secondary at 60 Hz is equivalent to 40V on the 16 Ohm winding at 20 Hz, for 100 Watts output. This gave 195 ma magnetizing current at the 16 Ohm winding, or equivalently 12 mA on the 4200 Ohm winding.
This means that 6 mA DC would use up half the core flux (on 4200 Ohm winding) for maximum symmetric signal flux, 0 to 12 mA. This 6 mA would also be the tube idle current.
So we then have 6 mA up or down from there for signal use. 6mA on the 4200 Ohm winding would then lead to a mere 0.15 Watts worth of useable flux for signal transformation. (Watts = I squared R) From a 100 Watt OT!!!! 0.15 Watts!!!!
To get 25 Watts out would require 77 mA up or down from DC idle (4200 Ohms) (you really need that gap to get 77 mA)
Looks like this P-P OT for SE idea is a really BAD JOKE. You would need headphones to hear anything out of a 100 Watt OT.
What the demo did was probably push the saturation current way up by a factor of 13 above that used by the xfmr manufacturer. So the thing would be operating so saturated, it would be like an air core xfmr. Band pass would start up around 260 Hz, like a telephone.
120V on the 16 Ohm secondary at 60 Hz is equivalent to 40V on the 16 Ohm winding at 20 Hz, for 100 Watts output. This gave 195 ma magnetizing current at the 16 Ohm winding, or equivalently 12 mA on the 4200 Ohm winding.
This means that 6 mA DC would use up half the core flux (on 4200 Ohm winding) for maximum symmetric signal flux, 0 to 12 mA. This 6 mA would also be the tube idle current.
So we then have 6 mA up or down from there for signal use. 6mA on the 4200 Ohm winding would then lead to a mere 0.15 Watts worth of useable flux for signal transformation. (Watts = I squared R) From a 100 Watt OT!!!! 0.15 Watts!!!!
To get 25 Watts out would require 77 mA up or down from DC idle (4200 Ohms) (you really need that gap to get 77 mA)
Looks like this P-P OT for SE idea is a really BAD JOKE. You would need headphones to hear anything out of a 100 Watt OT.
What the demo did was probably push the saturation current way up by a factor of 13 above that used by the xfmr manufacturer. So the thing would be operating so saturated, it would be like an air core xfmr. Band pass would start up around 260 Hz, like a telephone.
Last edited:
I’m guessing there’s a critical piece of info I have not yet discovered for how these amps are working. Parafeed seems the most logical but I cannot find the parts on the chassis that implement it. If I do run across some 845 tubes I will try these amps out. I’m not going to disassemble them anytime soon.
The thread I was thinking of was on AudioAsylum, mentioning amps by John Hogan:
Tube DIY Asylum
A quote from that was "John Hogan made these odd SE amps using the same PP opts for 2A3, 300B and 45s. They sounded good to people that heard them and he never argued about people questioning the application. He didn't know why they sounded good, they just did."
Although smoking-amps results seem to de-bunk this 😀
Tube DIY Asylum
A quote from that was "John Hogan made these odd SE amps using the same PP opts for 2A3, 300B and 45s. They sounded good to people that heard them and he never argued about people questioning the application. He didn't know why they sounded good, they just did."
Although smoking-amps results seem to de-bunk this 😀
Found the bit I was looking for. From Dave Slagle, a respected transformer winder:
".. found that a Dynaco A470 output had rapidly dropping inductance up to about 20ma where the L went from 180hy to around 20hy's. After 20ma the inductance dropped to 10hy @ 60ma and 7hy @ 100ma. This tells me that beyond the initial local saturation the thing actually becomes rather linear."
".. found that a Dynaco A470 output had rapidly dropping inductance up to about 20ma where the L went from 180hy to around 20hy's. After 20ma the inductance dropped to 10hy @ 60ma and 7hy @ 100ma. This tells me that beyond the initial local saturation the thing actually becomes rather linear."
I'm reaching way back in time here, but when the popularity of SE amps took off in the Sound Practices day, I heard of people disassembling and restacking Dynaco 70 A470 transformers to use single-ended amps. Long time ago and searches don't bring up much.
I once used Edcor 60W 4K4 transformer for 813 SE amp in class AB2, 20W with 800V B+.
I re-assembled it with an air gap. The result was great, due to interleaving.
I re-assembled it with an air gap. The result was great, due to interleaving.
Did anyone consider that maybe someone ordered the incorrect OT and then threw this project on the back burner...????
I did that with some of my surplus guitar amp OPT's. Yank out all the lams, stuff all the E's through the core, create the "gap" with a single layer of masking tape across the ends of the E's, stack up all the I's and place them in one end bell, put the core and E's in the same end bell, use a big C clamp to gently squeeze everything together, drop on the other end bell, put the bolts in and crank them tight since that's what keeps the gap from changing.
The Schumaker "80VA" 6600 ohm P-P OPTs would do 50 or so watts at 30 Hz, and over 150 watts at 1KHz in a push pull amp. The would just barely work with a 45 or triode wired 6V6 in SE before the restack. After restack I would call them a 10 watt OPT......maybe. Best driven with a triode wired KT88.
Outer_Space_10.....use a solid state CCS to stuff a sizable DC current through the secondary in the opposite direction of the current flowing in the primary. The current should be turns ratio times larger than the tube's idle current. Tried it, it works, but does put some DC on the speaker leads. Not a problem for some inefficient woofers, but don't do this if you are using 105db Lowthers.
----Correction---- to my Watts calculation for the P-P Edcor 4.2K when used for single ended.
I incorrectly used I squared R to calculate 0.15 Watts. But the flux restriction (to obey mfgrs max saturation spec) only restricts V, not I (same size wire).
So the max power out in SE (using mfgrs. Bmax for Sat.) would actually be 2 Watts. Still disappointing for a 100 Watt OT. Pushing Bmax up further will allow more power (more AC Volts), but also reduce the LF response. By the time it reaches 25 Watts, the LF response would rise to around 260 Hz, like an air core OT.
---------------------------------------------------------------------------------------
Thinking of high tech solutions, the Tektronix AM503/A5302 AC/DC current probe used a Hall effect device in the probe to sense DC magnetic fields. The AM503 adapter held an Op Amp which fed back DC into the probe to cancel the DC field. A DC servo in effect. A similar idea could be used to null the DC in a non gapped OT. Requires a Hall sensor in the OT and a controlled CCS to null the measured DC and a winding to apply the DC current.
I incorrectly used I squared R to calculate 0.15 Watts. But the flux restriction (to obey mfgrs max saturation spec) only restricts V, not I (same size wire).
So the max power out in SE (using mfgrs. Bmax for Sat.) would actually be 2 Watts. Still disappointing for a 100 Watt OT. Pushing Bmax up further will allow more power (more AC Volts), but also reduce the LF response. By the time it reaches 25 Watts, the LF response would rise to around 260 Hz, like an air core OT.
---------------------------------------------------------------------------------------
Thinking of high tech solutions, the Tektronix AM503/A5302 AC/DC current probe used a Hall effect device in the probe to sense DC magnetic fields. The AM503 adapter held an Op Amp which fed back DC into the probe to cancel the DC field. A DC servo in effect. A similar idea could be used to null the DC in a non gapped OT. Requires a Hall sensor in the OT and a controlled CCS to null the measured DC and a winding to apply the DC current.
Last edited:
We should never discount simple incompetence. Just look at historical battles or modern politics.
All good fortune,
Chris
- Status
- Not open for further replies.