I don't think it's correct the schem that you posted.I'm not sure though.But it isn't like McIntosh' unity coupled.
Anyway for a SE amp,you can't do it like McIntosh.But you can do it like Quad!
It's called CFB(cathode feedback).Do a google search to find the original Quad schem to see how it looks like.It's a PP design but...the half ... is a SE design!
You need a separate winding for the cathode.The only disadvantage is that you will need more preamp gain to drive the output stage.But the main advantage is that with CFB you can drive loads down to 2-3ohms!
Anyway for a SE amp,you can't do it like McIntosh.But you can do it like Quad!
It's called CFB(cathode feedback).Do a google search to find the original Quad schem to see how it looks like.It's a PP design but...the half ... is a SE design!
You need a separate winding for the cathode.The only disadvantage is that you will need more preamp gain to drive the output stage.But the main advantage is that with CFB you can drive loads down to 2-3ohms!
The schematic you show seems to be a simple SEUL (single ended ultra-linear) amplifier. Since a triode does not have a screen grid, you cannot use it un UL connection this way. Have a look at Patrick Turner's 22W SEUL 13E1 amp
Yes, you can use cathode feedback taps on a single ended amp to good effect. Have a look at the "Scrapbox Challenge" amplifier in Valve Amplifiers if you have a copy.
Yes, you can use cathode feedback taps on a single ended amp to good effect. Have a look at the "Scrapbox Challenge" amplifier in Valve Amplifiers if you have a copy.
On second thoughts, I probably should add that using cathode feedback on the output stage of an amp by leaving Rk unbypassed isn't a great idea when using anything in fixed screen mode, because the valve will increase its effective anode resistace by Rk x m (which is large), so you end up driving the OPT with what is basically a CCS, which isn't all that great for bandwidth or the damping factor of the amp. Somewhat less of a problem with UL or triode connection, but CFB via taps on the transformer is a better solution, if you can get them since CFB this way reduces output impedance.
resident is speaking mcintosh and planet10 quad for me
it s a good way to go
for more feedback :
-speaker to 8ohm output
-cathode feedback to 16 ohm output
a this time nobody bring a working se amp with standard transfo
in all internet maybe it s a door to open for SE amp
i hope this discussion find a pratical key for SE amp
it s a good way to go
for more feedback :
-speaker to 8ohm output
-cathode feedback to 16 ohm output
a this time nobody bring a working se amp with standard transfo
in all internet maybe it s a door to open for SE amp
i hope this discussion find a pratical key for SE amp
yes yes silurato i m very interested
my email
jeapel@loginnovation.com
my internet site
www.loginnovation.com/jeapel
thank you very much
silurato i m very interested my email
jeapel@loginnovation.com
my internet site
www.loginnovation.com/jeapel
thank you very much
From lundahl transfo web site a cathode feedback
with normal transfo
http://www.lundahl.se/claus_b_se.html
with normal transfo
http://www.lundahl.se/claus_b_se.html
There's no reason why you can't use cathode feedback on a single ended output stage. Note that it will be similar, but not identical to McIntosh's UnityCoupled output topology (only works for push-pull, and with special transformers).
The schmatic as you have drawn it should be good. I would bypass the cathode, if I were you. And you can experiment with connecting the screen in either pentode, triode, or UltraLinear mode. What tube are you using? 6L6? EL34? the KT__ gang?
take a look at the reverb drive stage in this Fender Reverb Unit (from schematicheaven.com) https://schematicheavencom.secure.powweb.com/fenderamps/reverb_sf_schem.pdf
This shows how to apply cathode feedback with an ordinary transformer, if you bias your tube using a fixed negative bias (derived via "back bias" or the traditional bias supply). If you want to use cathode bias, you can either do it like your schematic (speaker grounded), or you can connect the cathode directly to the output winding, and then the ground of the winding goes to the cathode resistor/cap (speaker floating at cathode potential, harder to take "standard" global negative feedback, not recommended).
I am currently using a small stereo push-pull EL84 amplifier for my computer sounds. It is using the "trebble" channel outputs transformers from some Hammond PR-40 "tone cabinet" church organ speaker (overbuilt like crazy, the "trebble" trannies are big enough for normal full-range operation. I'd say about as big as a 2xEL84 guitar amp tranny, or bigger than a universal hammond tranny.). The transformers are like regular 8K to 8 ohm p-p but with a seperate 40% screen tap. I have a regular DPDT input select switch, a 100K volume control, a 12AX7 voltage gain stage, which is capacitively coupled to a 12AU7 split load phase inverter. The preamp-driver is similar to an Eico HF-86, but I didn't directly connect the phase splitter's grid to the coltage gain stage's plate. I used a coupling cap there, and then used a voltage divider from B+ to ground (1meg, 470K) to bias the phase inverter. I considered adding in global negative feedback (to the cathode of the voltage gain stage), but the amplifier seemed perfectly fine with just cathode feedback for the output stage. It does MORE than i need for my smallish computer speakers. I can also connect my 4 ohm Dayton D-III's to it, and play AC-DC (Church Organ parts don't object to the devil's music?) at the...correct...volume level for AC/DC with pounding bass. The amplifier doesn't struggle at all like this. I'm a big fan of cathode feedback now
The schmatic as you have drawn it should be good. I would bypass the cathode, if I were you. And you can experiment with connecting the screen in either pentode, triode, or UltraLinear mode. What tube are you using? 6L6? EL34? the KT__ gang?
take a look at the reverb drive stage in this Fender Reverb Unit (from schematicheaven.com) https://schematicheavencom.secure.powweb.com/fenderamps/reverb_sf_schem.pdf
This shows how to apply cathode feedback with an ordinary transformer, if you bias your tube using a fixed negative bias (derived via "back bias" or the traditional bias supply). If you want to use cathode bias, you can either do it like your schematic (speaker grounded), or you can connect the cathode directly to the output winding, and then the ground of the winding goes to the cathode resistor/cap (speaker floating at cathode potential, harder to take "standard" global negative feedback, not recommended).
I am currently using a small stereo push-pull EL84 amplifier for my computer sounds. It is using the "trebble" channel outputs transformers from some Hammond PR-40 "tone cabinet" church organ speaker (overbuilt like crazy, the "trebble" trannies are big enough for normal full-range operation. I'd say about as big as a 2xEL84 guitar amp tranny, or bigger than a universal hammond tranny.). The transformers are like regular 8K to 8 ohm p-p but with a seperate 40% screen tap. I have a regular DPDT input select switch, a 100K volume control, a 12AX7 voltage gain stage, which is capacitively coupled to a 12AU7 split load phase inverter. The preamp-driver is similar to an Eico HF-86, but I didn't directly connect the phase splitter's grid to the coltage gain stage's plate. I used a coupling cap there, and then used a voltage divider from B+ to ground (1meg, 470K) to bias the phase inverter. I considered adding in global negative feedback (to the cathode of the voltage gain stage), but the amplifier seemed perfectly fine with just cathode feedback for the output stage. It does MORE than i need for my smallish computer speakers. I can also connect my 4 ohm Dayton D-III's to it, and play AC-DC (Church Organ parts don't object to the devil's music?) at the...correct...volume level for AC/DC with pounding bass. The amplifier doesn't struggle at all like this. I'm a big fan of cathode feedback now
> The vintage mcintosh amp use a separate winding on transfo output to create a local cathode feedback. But for a SE amp with normal transfo is it possible ? and a good idea ? to do that.
Possible? Sure. Good idea? Probably a waste of time, especially for Triode.
The Mac is "unity coupled": the cathode swings as much as the plate. If the B+ is 450V, and the tube can suck down to 50V Plate-Cathode, a normal amp will swing the plate 50V-850V. The Mac will swing half of this on the plate, 250V-650V, and half on the cathode: +200V to -200V. The grid-cathode swing for the 7195 and 6L6 tubes Mac used is 25V-50V, so the total grid swing is large compared to grid-cathode swing. The tube works with very large feedback.
But a 3K:8 ohm transformer is a very different thing. The winding ratio is closer to 20:1 than 1:1.
Consider a basic SET with 2A3, working at book values 250V, -45V, 60mA, 2K5 load, 3.5 Watts.
A 2K5:16 transformer is available (probably). 3.5 Watts in 16 ohms is 7.5V RMS, 10.6V peak. The grid-cathode swing is 45V peak. So sticking the 16 ohm winding under the cathode increases total grid swing from 45V to 55.6V. The amount of feedback is proportional to the reduction in gain. Gain is reduced less than 2dB, so we don't even have 2dB of feedback. That is nearly insignificant, and I suspect the side effects are not for the good.
Also, raising the drive from 45V to 55.6V increases the burden on the driver, which (if it is R-C coupled and running on the same supply as the 2A3) is already working as hard as it can.
The problem is the low load impedances we like for speakers (limited by wire gauge and market habit) and the very low voltage gain of a high-current triode. Going to a high-Mu triode seems better, but the maximum current (for similar heater power) is less. No improvement.
Things are "better" with the high-gain Pentodes. 8417 will make around 15 Watts with about 15V RMS drive. With a little rounding, this is 20V peak drive, and 20V peak in the 16 ohm winding. Drive voltage increases from 20V to 40V, and you really do have 6dB of feedback, enough to take some of the curse off of an SEP power amp. And the transformer is about the only significant reactance in the feedback loop. Stability is solid. 40V peak drive is not a major problem. 6dB feedback is enough to give some damping and a little better response smoothness. But it also puts output distortion back into the tube to be intermodulated. With complex signal this makes low-level hash in the background, and it may take another 30dB NFB to reduce this to the level of no NFB. While many low-NFB amps are happy things, I'm inclined to either no NFB or a LOT of NFB, not 6dB NFB.
In a sense, what transformer cathode coupling does is change the effective Mu of the tube without (much) changing its current capacity. With a 20:1 transformer ratio, a 2A3's Mu can't be over 20, but is really only 4, so effective Mu is hardly changed. With a Mu of 20 and a transformer forcing effective Mu no higher than 20, we do get some benefit. But Mu=20 triodes don't make as much current as low-Mu triodes, won't make much power. Pentodes are different, But with speaker-impedance windings and typical amplifier powers, we only get a little NFB. You have to go to much lower winding ratios (higher winding impedance) to get a lot of improvement. McIntosh amps force Mu~2 on tubes that act like Mu is around 20, so are high NFB, but special transformers.
If you are driving 600 ohm outputs, things are better. But very often when we want "600 ohms", we would also like "floating output" to reduce problems on long lines. And the output can't drive a cathode and float too.
Yet another "cathode feedback" trick is an old TEAC tapedeck headphone output. Single transistor with 1K winding in the collector and 8 ohm winding in the emitter. It acts like an emitter follower, but the output current can be much higher than the supply current due to the 1K:8 transformation in the collector. The Fender Reverb driver Photon points to has similar transformation.
> with CFB you can drive loads down to 2-3ohms!
This plan does not change the large-signal power relationships, except for needing increased drive signal. (A small correction for transformer ratio, but for typical speaker transformers with 20:1 or 30:1 ratio this is negligible.)
Possible? Sure. Good idea? Probably a waste of time, especially for Triode.
The Mac is "unity coupled": the cathode swings as much as the plate. If the B+ is 450V, and the tube can suck down to 50V Plate-Cathode, a normal amp will swing the plate 50V-850V. The Mac will swing half of this on the plate, 250V-650V, and half on the cathode: +200V to -200V. The grid-cathode swing for the 7195 and 6L6 tubes Mac used is 25V-50V, so the total grid swing is large compared to grid-cathode swing. The tube works with very large feedback.
But a 3K:8 ohm transformer is a very different thing. The winding ratio is closer to 20:1 than 1:1.
Consider a basic SET with 2A3, working at book values 250V, -45V, 60mA, 2K5 load, 3.5 Watts.
A 2K5:16 transformer is available (probably). 3.5 Watts in 16 ohms is 7.5V RMS, 10.6V peak. The grid-cathode swing is 45V peak. So sticking the 16 ohm winding under the cathode increases total grid swing from 45V to 55.6V. The amount of feedback is proportional to the reduction in gain. Gain is reduced less than 2dB, so we don't even have 2dB of feedback. That is nearly insignificant, and I suspect the side effects are not for the good.
Also, raising the drive from 45V to 55.6V increases the burden on the driver, which (if it is R-C coupled and running on the same supply as the 2A3) is already working as hard as it can.
The problem is the low load impedances we like for speakers (limited by wire gauge and market habit) and the very low voltage gain of a high-current triode. Going to a high-Mu triode seems better, but the maximum current (for similar heater power) is less. No improvement.
Things are "better" with the high-gain Pentodes. 8417 will make around 15 Watts with about 15V RMS drive. With a little rounding, this is 20V peak drive, and 20V peak in the 16 ohm winding. Drive voltage increases from 20V to 40V, and you really do have 6dB of feedback, enough to take some of the curse off of an SEP power amp. And the transformer is about the only significant reactance in the feedback loop. Stability is solid. 40V peak drive is not a major problem. 6dB feedback is enough to give some damping and a little better response smoothness. But it also puts output distortion back into the tube to be intermodulated. With complex signal this makes low-level hash in the background, and it may take another 30dB NFB to reduce this to the level of no NFB. While many low-NFB amps are happy things, I'm inclined to either no NFB or a LOT of NFB, not 6dB NFB.
In a sense, what transformer cathode coupling does is change the effective Mu of the tube without (much) changing its current capacity. With a 20:1 transformer ratio, a 2A3's Mu can't be over 20, but is really only 4, so effective Mu is hardly changed. With a Mu of 20 and a transformer forcing effective Mu no higher than 20, we do get some benefit. But Mu=20 triodes don't make as much current as low-Mu triodes, won't make much power. Pentodes are different, But with speaker-impedance windings and typical amplifier powers, we only get a little NFB. You have to go to much lower winding ratios (higher winding impedance) to get a lot of improvement. McIntosh amps force Mu~2 on tubes that act like Mu is around 20, so are high NFB, but special transformers.
If you are driving 600 ohm outputs, things are better. But very often when we want "600 ohms", we would also like "floating output" to reduce problems on long lines. And the output can't drive a cathode and float too.
Yet another "cathode feedback" trick is an old TEAC tapedeck headphone output. Single transistor with 1K winding in the collector and 8 ohm winding in the emitter. It acts like an emitter follower, but the output current can be much higher than the supply current due to the 1K:8 transformation in the collector. The Fender Reverb driver Photon points to has similar transformation.
> with CFB you can drive loads down to 2-3ohms!
This plan does not change the large-signal power relationships, except for needing increased drive signal. (A small correction for transformer ratio, but for typical speaker transformers with 20:1 or 30:1 ratio this is negligible.)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.