Ah, yes. I've been interested in them for a driver also. I want to try a scheme where local N Fdbk goes back to the 12HL7 screen grid so as to make the actual output tube act as a triode equivalent to the 12HL7 internal triode. Jan Veiset came up with idea a long while back. If the Fdbk matches the internal triode mu, then the driver current stays steady. But if not, the drive adjusts so as to get it to match. "Tube emulation" essentially. Then I can make all those cheap TV Sweeps play like a 211.
Looks like triode mu for 6197 is in the 22 to 25 range, depending on current. I only have GE ones.
Triode mu for the 12HL7 I have seen between 25 and 40, depending on manufacturer. RCA 40, GE 30, SYL 25 and current somewhat. (limited # I have tested)
Triode mu for the 12HL7 I have seen between 25 and 40, depending on manufacturer. RCA 40, GE 30, SYL 25 and current somewhat. (limited # I have tested)
Last edited:
What screen voltage was used for that trace? I'm using a 12gn7 pentode-mode as an 813 driver but find them moderately microphonic.
The 6197 pentode curves were done with 110V on grid 2. 50V/div. Horizontal and 5mA/div Vertical, with 0.25V grid1 steps.
I looked at some tubes here,
GE 6197,
GE 12HG7, Lindal, Zenith, Sylvania & RCA 12HL7,
RCA & GE 6HB6, RCA 6AH9 (contains 12HL7 pentode ) .
Only the 6AH9 (12HL7 pent. inside) had a double mica, but just on the top end. The Lindal and Sylvania 12HL7 had either thicker or fluffed up bottom micas. Found another Sylvania 12HL7 with just single micas though. The GE 12HG7 had a thicker bottom mica.
The GE 6197 does have metal plates covering most of both micas, so seems to have some extra measure of rigidity built in.
Some other frame grid tubes are the 9DX base tubes. Found a Toshiba 10KR8, a Westinghouse 11KV8, Sylvania 6LY8 and Lindal 6JT8, and all of them had either a thicker or doubled bottom mica.
And a CCCP 6F12P here has either doubled or thicker micas both top and bottom.
For microphonic susceptible tubes, I would suggest tapping them and listening for a ring tone. Some of the above do produce a tone.
Last edited:
All this has gone a bit far afield from the original question, which referred to the virtues (dubious or not) of the newer Hammond 1628SEA as compared to the older version. Tubelab has had a fair bit to say about it in earlier posts, mostly that the earlier version of the transformer was definitely deficient in terms of HF response. I don't remember what he had to say about its newer incarnation.
This thread had a discussion comparing similar Edcor and Hammond product:
https://www.diyaudio.com/community/threads/hammond-and-edcor-se-xfmrs-compared.387442/#post-7054525
This thread had a discussion comparing similar Edcor and Hammond product:
https://www.diyaudio.com/community/threads/hammond-and-edcor-se-xfmrs-compared.387442/#post-7054525
Last edited:
😳 Oh yeah, 1628SEA. No micas there.
I would guess that Hammond fixed the 1628SE dropout problem with their whole new series. Dropping the primary inductance is the right general fix (beside any winding problem). But I sure would check that new 10K Z SE OT with a swept osc. and scope.
I would guess that Hammond fixed the 1628SE dropout problem with their whole new series. Dropping the primary inductance is the right general fix (beside any winding problem). But I sure would check that new 10K Z SE OT with a swept osc. and scope.
Last edited:
Me, I'm lugging a pair of Sansui canned and potted P-P output transformers (acquired relatively on the cheap, as compared to their far more expensive Hashimoto descendants) in to my workplace for characterization. They were probably designed for use with a pair of 6BQ5s/EL84s. I'll winkle out their secrets on the bench, including what mounting screws were intended for use with them - I suspect they might take metric screws. They will definitely get stripped with a wire wheel and acquire a new coat of paint, as the current finish is pretty dull and ugly.
Last edited:
Metric screws always measure nicely; whole numbers.
Inch screws are most always a fraction of what they could be.
In 1960, the US government declared that the US was going to go all Metric.
My, they were very Prophetic.
We are finally starting to follow the rest of the world, we will get there eventually.
If you find you do not like the Sansui transformers, let me know. Ha Ha
I might purchase them.
EL84 or if they are big, 7591 perhaps.
Inch screws are most always a fraction of what they could be.
In 1960, the US government declared that the US was going to go all Metric.
My, they were very Prophetic.
We are finally starting to follow the rest of the world, we will get there eventually.
If you find you do not like the Sansui transformers, let me know. Ha Ha
I might purchase them.
EL84 or if they are big, 7591 perhaps.
Last edited:
I think the Ebay seller referred to them as EL84 output iron - they are kinda on the smallish side. I'll be hanging on to them, as I haven't finished an EL84 output-based design yet. I have one in the works with a hybrid front end, but it got lost in the shuffle and is buried under a pile of other projects.
Geez, I can't believe Edcor's 25 Watt P-P OTs are $112 now. 25 Watt SE OTs are $148. And 15 Watt are $81 - $92.
So much for "small-little" projects.
So much for "small-little" projects.
Last edited:
Yes, I was also shocked by Edcor's sudden pricing bloat - maybe they are getting hit hard for sourcing the basic materials. Just like Hammond, they used to be a bargain, but no more...
The original 1628SE just sucked, period. I couldn't even make a decent sounding guitar amp with them. I have a pair of 1628SEAs that wear a 2006 build date on the label. I also have a pair of the big 25 watt 5K ohm Edcors from the same time period. In everything I have tried both in, the Edcors always produce 5 to 10% more power than the Hammonds in the same circuit. The Hammonds have the best bass in both measurements and listening tests, while the Edcors have the lead in high end response. The Hammonds have never been mounted in a complete amp, and the Edcors spent about a year in a 300B amp before being voted off the top plate by a set of Transcendars. Both sets have been test mules ever since.
OK, the Hammonds have more iron and copper, so this is as it should be. If you plan to drive either big OPT with a wimpy tube and don't put any feedback around it to lower its plate resistance in order to shunt the effects caused by the leakage inductance resonating with the stray capacitance to form a notch you will lose some high end. Remember that the "Q" of a resonant circuit diminishes as the resistance in parallel with it drops. Drive those big boys with a very low impedance source if you want them to play nice.
This may change if I get back to making HiFi amps. I was doing some testing when I created a SE test amp with two FAT TV sweep tubes in parallel wired UNSET configuration and I finally got to hear those big Hammonds sing. They passed almost 40 watts at 30 Hz too! The Edcors were a no show in this event since I was running the Hammonds at 2500 ohms by connecting my 8 ohm speakers (or load resistor) to the 16 ohm tap. The Edcors only have an 8 ohm output.
OK, the Hammonds have more iron and copper, so this is as it should be. If you plan to drive either big OPT with a wimpy tube and don't put any feedback around it to lower its plate resistance in order to shunt the effects caused by the leakage inductance resonating with the stray capacitance to form a notch you will lose some high end. Remember that the "Q" of a resonant circuit diminishes as the resistance in parallel with it drops. Drive those big boys with a very low impedance source if you want them to play nice.
This may change if I get back to making HiFi amps. I was doing some testing when I created a SE test amp with two FAT TV sweep tubes in parallel wired UNSET configuration and I finally got to hear those big Hammonds sing. They passed almost 40 watts at 30 Hz too! The Edcors were a no show in this event since I was running the Hammonds at 2500 ohms by connecting my 8 ohm speakers (or load resistor) to the 16 ohm tap. The Edcors only have an 8 ohm output.
I still have the 1628SE for bench beaters. Just measured >400 ohms primary DCR. In their current application that's over 40 volts B+ loss so not surprising. That said, off the 4 ohm tap and wrapped in 20 dB of very low distortion feedback they measure and sound pretty good. 20 watts at 30 Hz well under 1% distortion.
I get the sense that OTLs, hybrids, and "tube like" ss amps will have a resonance within the next couple years.
Either that or we will start seeing some more diy friendly winding machines. Either way, the economics of transformer winding is kind of upside down these days.
Umm, I forgot the primary resistance for the Hammond 1640SE (1250 Ohm Z ) earlier. Just checked it, got 93 Ohms DC.
Looks to be proportionate to primary Z. (400 Ohm DC for the 1638SE above, 5000 Ohm Z)
*****************************************************************************************************************************
I did get some interesting adjustable dual switching supplies, with linear post regulators, off Epray some years ago. Adjustable, 50V to 155V at 6 Amps for each internal supply unit. $30 Were built for NASA by Xantrex. With boxes of 21LG6s or 6HJ5s and some 6528 regulator triodes , I'm set to build most any OTLs, BUT I would use an OT now using the OTL cheat patent, US4614914. No heat wave,. Only two or four output tubes. Even have a couple of Toroidy 1200 Ohm P-P 80 W OTs available.
Another OT approach I looked into was using small power xfmrs in combination. Some years back I got a bunch of Signal Xfmr DPC-120-200 (dual 120V windings and dual 60V windings) and a bunch of Signal Xfmr DPC-40-600 (dual 120V windings and dual 20V windings), among others. All are 24 Watt at 60 Hz. Cost $3 each back then, but are $15 now! Trying to split a P-P primary up between multiple xfmrs gets into too much capacitance, but Circlotron mode looks good for them.
I think the real DIY answer is to develop a small PC board with a standard switchmode IC for square wave inverters (ie., Berning type impedance converter) (which includes crossover deadtime and clock), low and high side Mosfet drivers, and positions for 4 power Mosfet switches (going onto a heatsink). Fool-proof with the IC, no tickler Fdbk osc. You wind the modest # turns ferrite toroid xfmrs yourself, with some written advice on keeping leakage L minimized and winding capacitance minimized.
Looks to be proportionate to primary Z. (400 Ohm DC for the 1638SE above, 5000 Ohm Z)
*****************************************************************************************************************************
OTL, I still don't like 5% efficiency. Which means the power supply will be big and heavy and expensive. And the heat wave. Shipping cost lately is prohibitive too.
I did get some interesting adjustable dual switching supplies, with linear post regulators, off Epray some years ago. Adjustable, 50V to 155V at 6 Amps for each internal supply unit. $30 Were built for NASA by Xantrex. With boxes of 21LG6s or 6HJ5s and some 6528 regulator triodes , I'm set to build most any OTLs, BUT I would use an OT now using the OTL cheat patent, US4614914. No heat wave,. Only two or four output tubes. Even have a couple of Toroidy 1200 Ohm P-P 80 W OTs available.
Another OT approach I looked into was using small power xfmrs in combination. Some years back I got a bunch of Signal Xfmr DPC-120-200 (dual 120V windings and dual 60V windings) and a bunch of Signal Xfmr DPC-40-600 (dual 120V windings and dual 20V windings), among others. All are 24 Watt at 60 Hz. Cost $3 each back then, but are $15 now! Trying to split a P-P primary up between multiple xfmrs gets into too much capacitance, but Circlotron mode looks good for them.
I think the real DIY answer is to develop a small PC board with a standard switchmode IC for square wave inverters (ie., Berning type impedance converter) (which includes crossover deadtime and clock), low and high side Mosfet drivers, and positions for 4 power Mosfet switches (going onto a heatsink). Fool-proof with the IC, no tickler Fdbk osc. You wind the modest # turns ferrite toroid xfmrs yourself, with some written advice on keeping leakage L minimized and winding capacitance minimized.
Last edited:
DIY OTs (switchmode, high mu ferrite) Cheap, easy, customize-able, light weight
But I wonder if you can find suitable ferrite cores easily on Ebay now. Nothing made here anymore. Well, there is Fair-Rite products still and Magnetics Inc.
But I wonder if you can find suitable ferrite cores easily on Ebay now. Nothing made here anymore. Well, there is Fair-Rite products still and Magnetics Inc.
Last edited:
Suitable cores (and bobbins) are easily found at Digi-Key and Mouser. Dealing with the ferrites available on E-pray is not the best option, as you're not entirely sure what you're getting in terms of material. With the distributors, you can pick a suitable ferrite and be sure of what you're getting.
Last edited: