Fortunately, the design I'm working with features the 6BM8 in single-ended mode rather than push-pull. The output power will be less (likely dramatically so), but maybe still sufficient to drive my Klipsch speakers to audible output levels.
Just for the sake of completeness, I need to find one of my old Edcor XSE 5k output transformers and test it on the bench. I'm curious as to the results I will get for turns ratio accuracy and primary inductance. I remember them being deficient with my "Shrine" SE amp design, with pronounced tilt when driven by a low frequency square wave. I replaced them with GXSE transformers (just introduced, and with open-frame construction rather than end bells), and got much better bass response.
Last edited:
Single ended output transformers that are driven with a low frequency square wave (like 100Hz square wave) . . .
an oscilloscope will show the Tilt that is present.
Additionally, if low frequency sine waves, or low frequency square waves cause the transformer to go into saturation . . .
Global Negative Feedback will only make the transformer even more saturated.
Test turns ratio: use a low impedance to drive the primary at 1kHz. Do not put a load on the secondary.
Measure the ratio of the primary voltage Vp / secondary voltage Vs.
Vp/Vs is the turns ratio.
When you use a low impedance signal source, it swamps out the primary inductance, and the distributed capacitance that are in parallel, and usually resonates somewhere between 500Hz and 2kHz. If you used a high impedance signal source, you will get the wrong VP/Vs ratio, and not know the True Turns ratio.
If the secondary has an 8 Ohms rating, then the primary impedance is 8 x (Vp/Vs)squared = primary Ohms
If the voltage ratio is 25, then Vp/Vs = 25, and the primary is 8 Ohms x (25 squared) = 8 x 625 = 5000 Ohms
When the 8 Ohm tap is loaded with an 8 Ohm resistor, there is an insertion loss of the transformer
The primary DCR is in series with the primary turns; and the secondary DCR is in series with the secondary turns.
With an 8 Ohm load on the 8 Ohm tap, the primary impedance will be More than 5k Ohms.
an oscilloscope will show the Tilt that is present.
Additionally, if low frequency sine waves, or low frequency square waves cause the transformer to go into saturation . . .
Global Negative Feedback will only make the transformer even more saturated.
Test turns ratio: use a low impedance to drive the primary at 1kHz. Do not put a load on the secondary.
Measure the ratio of the primary voltage Vp / secondary voltage Vs.
Vp/Vs is the turns ratio.
When you use a low impedance signal source, it swamps out the primary inductance, and the distributed capacitance that are in parallel, and usually resonates somewhere between 500Hz and 2kHz. If you used a high impedance signal source, you will get the wrong VP/Vs ratio, and not know the True Turns ratio.
If the secondary has an 8 Ohms rating, then the primary impedance is 8 x (Vp/Vs)squared = primary Ohms
If the voltage ratio is 25, then Vp/Vs = 25, and the primary is 8 Ohms x (25 squared) = 8 x 625 = 5000 Ohms
When the 8 Ohm tap is loaded with an 8 Ohm resistor, there is an insertion loss of the transformer
The primary DCR is in series with the primary turns; and the secondary DCR is in series with the secondary turns.
With an 8 Ohm load on the 8 Ohm tap, the primary impedance will be More than 5k Ohms.
Last edited:
There was a lot of tilt with the Edcor XSE transformers. That may mean lowish primary inductance. I will have to root one out from the mess on my "construction" table and throw it on my bench at work for tests. The Edcor GXSE transformers were much better bass-wise. I will need to run the amp again and see what sort of tilt happens with the 5k GXSE transformers. As with most of my projects, it's in the throes of a significant transformation, from SE Ultralinear to Schade-style feedback. That may require me to root out my Johnson square ceramic octal sockets which might fit on the same centers as the sockets for the original 1625 outputs.
Last edited:
Because I am a complete novice, I was hoping to find detailed build plans for a similar amp to the 6BM8, such as the ECL86 which is a bit more powerful. I am wondering if there was a way to "upgrade' ? I suspect that bias points would need to change etc.
R
R
The ECL86 is quite a bit more expensive than the 6BM8 - check Ebay and I think you'll get a bad case of sticker shock.
If you want more power from a "spud" amp, try this 6LU8, a Kegger inspired unit I also built. With the 15W GXSE Edor iron, this amp performs very close to what an EL34 amp would do.
https://www.skunkiedesigns.com/6lu8
https://www.skunkiedesigns.com/6lu8
I built a bitty - bitty amp using the 6LR8 with the pentode section trioded. It used a small flyback SMPS to power it. I should resurrect that amp, as I now have a set of pretty efficient Klipsch speakers in the living room that might play nicely with the lowish power available from the amp.
It's described in this thread:
https://www.diyaudio.com/community/threads/mighty-mite-single-ended-amplifier.120804/
It's described in this thread:
https://www.diyaudio.com/community/threads/mighty-mite-single-ended-amplifier.120804/
Last edited:
6MF8/15MF8 is another compactron that has spud potential and it's pin compatible with 6LU8 if memory serves. 6HZ8 is your most over powered option for triode-pentode tube if you must stay with 9-pin miniature sockets. Make sure you have enough heater supply current though. The best part, it's in the 9DX pin layout family - shared with a bunch of other triode-pentode tubes. Another 9DX tube that sim's above it's weight is 6GN8/6EB8. A perfect Shade feedback candidate with its high gain triode. I've looked at a lot of these for PP duty using toroidal transformers as OP iron. A lot of them look promising in simulations.
The next one I do may be similar to to a headphone amp I did using a depletion mode mosfet as front end and maybe something like a 6BQ5 or 6V6 for the output, partial feedback "Schade" style (I actually used a 12AQ5 as output for the headphone amp). I may go with a pair of the 5k "H'Fi" transformers from Musical Power Supplies that look to have specs similar to the GXSE from Edcor, but with 10W capability.
Make sure you get their "HiFi" transformers, they are quite good. Their cheaper normal guitar amp models aren't that great for HiFi use.
If you dont need as much power as the LU/LR8 and want 9 pin mini there is the 6KV8. The pentode is a high gain video output (which are usually very good trioded). The small signal triode is also quite good like the LU8’s is. There is also the KR8 which is the same micro power pentode, with a triode with lots of 2HD. Got a guitar amp in the works using only 6KR8’s. And you can guess what it’s called….
If you are looking at other candidates, not sure if you mentioned it, but there is also 6GV8/ECL85. There is also very similar ECL805, and serial heater version PCL85/PLC805. They are still all cheap. I have a B&O amplifier that uses ECL85 for output.
Pye used PCL83 in an amplifier. Then there are the 10 pin tubes, like ECL200 ...
Most of these don't have American style designations. Did Europe go its own way at some point?
So many candidates, so little time.
Pye used PCL83 in an amplifier. Then there are the 10 pin tubes, like ECL200 ...
Most of these don't have American style designations. Did Europe go its own way at some point?
So many candidates, so little time.
I built the 6LU8, it is very good. It is my main music listening system. They drive my 89db sensitive speakers to quite good listening level in my 16x19 media room..
When I was looking for my first project, I wanted to try an ECL86 but was unable to find as detailed a build guide as Stephek's series. I am very happy with the 6BM8 amp. For use at my desk, with smallnearfieldspeakers, is sounds great.
Roger
When I was looking for my first project, I wanted to try an ECL86 but was unable to find as detailed a build guide as Stephek's series. I am very happy with the 6BM8 amp. For use at my desk, with smallnearfieldspeakers, is sounds great.
Roger
wrenchone,
Test square wave tilt?
The very old rule of thumb was a good looking 1kHz square wave meant the amplifier was OK from 100Hz to 10kHz.
Was there an actual measurement of bandwidth this way, No. Not without doing some calculations.
I use my Denon Audio Technical CD to do spot frequency response checks.
21.5Hz, 40Hz, 100Hz, 1kHz, 9,999Hz, and 19,999Hz.
These are all equal amplitude sine waves.
20 log (V test frequency / V 1kHz) = dB (usually a negative number at any frequency other than the 1kHz reference frequency).
Just a spot frequency response test.
But remember:
If a low frequency response is -1dB at a given frequency, the -3dB is at 1/2 that frequency (with a single pole response);
And
If a high frequency response is - 1dB at a given frequency, there will be -3dB or more attenuation at 2 x that frequency (with a 2 pole response; which falls more quickly than a single pole roll off).
I also use the Denon Audio Technical CD 'Test Impulse' to do a quick and easy look at the bandwidth up to 22.05kHz.
The oscilloscope FFT shows the frequency response.
I only use a fast-rise fast-fall square wave to test 2 things:
1. The amplifier bandwidth (0.35 / rise-time; and 0.35 / fall-time) (a CD player can not be used to test bandwidth higher than 22.05kHz)
2. To look for any high frequency "ringing". Resonance, not oscillation. My amplifiers do not oscillate.
A 100Hz square wave applied to a single ended amplifier that does not have global negative feedback, has lots of tilt.
No, I am not suggesting using global negative feedback there, because . . .
If there is any saturation of the 100Hz square wave, then global negative feedback makes the saturation even worse.
Have fun testing the Edcor transformers again.
Test square wave tilt?
The very old rule of thumb was a good looking 1kHz square wave meant the amplifier was OK from 100Hz to 10kHz.
Was there an actual measurement of bandwidth this way, No. Not without doing some calculations.
I use my Denon Audio Technical CD to do spot frequency response checks.
21.5Hz, 40Hz, 100Hz, 1kHz, 9,999Hz, and 19,999Hz.
These are all equal amplitude sine waves.
20 log (V test frequency / V 1kHz) = dB (usually a negative number at any frequency other than the 1kHz reference frequency).
Just a spot frequency response test.
But remember:
If a low frequency response is -1dB at a given frequency, the -3dB is at 1/2 that frequency (with a single pole response);
And
If a high frequency response is - 1dB at a given frequency, there will be -3dB or more attenuation at 2 x that frequency (with a 2 pole response; which falls more quickly than a single pole roll off).
I also use the Denon Audio Technical CD 'Test Impulse' to do a quick and easy look at the bandwidth up to 22.05kHz.
The oscilloscope FFT shows the frequency response.
I only use a fast-rise fast-fall square wave to test 2 things:
1. The amplifier bandwidth (0.35 / rise-time; and 0.35 / fall-time) (a CD player can not be used to test bandwidth higher than 22.05kHz)
2. To look for any high frequency "ringing". Resonance, not oscillation. My amplifiers do not oscillate.
A 100Hz square wave applied to a single ended amplifier that does not have global negative feedback, has lots of tilt.
No, I am not suggesting using global negative feedback there, because . . .
If there is any saturation of the 100Hz square wave, then global negative feedback makes the saturation even worse.
Have fun testing the Edcor transformers again.
Last edited:
I'm not using the Edcor transformers (only 10H primary inductance). I'm going instead with a pair looted out of an old Sony tape deck that used SE 6AQ5s for outputs. Those have a respectable 20-21 H primary inductance and a higher current rating than the wimpy Edcors, which were only rated for 40 mA. The turns ratio on the two tests out at 4500-47000 :8. I will be running tests on this amp with a bona-fide gain-phase analyzer to winkle out all its little kinks and secrets
Oh, and I found my "Mighty-Mite" SE amp in the pile - I will be fitting it with a new set of Sears 6LR8s and firing it up, hopefully without any fire involved. The Sears 6LR8s look to be re-marked Sylvanias. I may also try looting a pair of GEs from a hulk that I have downstairs. The Sylvanias have the neat tip-off on top, while the GEs have smooth tops. The amp used Hammond 125CSEs for output transformers, set up for 5k:8 primary impedance.
I brought the recovered Mighty-Mite SE amp up to work and re-fitted it with a set of Sears 6LR8s, which look to be re-labelled Sylvanias - they have the cool tip-off on top. The amp fired up without bursting into flame, though the big pentodes look to be getting pretty hot, run hard in triode mode. The amp sounded OK on it's last sojurn in my living room - I'll see how loud it gets running a pair of pretty efficient Klipsch (Tangent series) floor-standing speakers. It could be a fun little amp.
Which Tangent model do you have? I recently scored a good deal on a pair of 400's. They're basically a Heresy II in larger ported(albeit lower build quality) cabinets. Do you plan on doing any upgrades?