I think the title says it all. I want to tinker with parafeed, and I like EL84's, but the 40K-ish Ra is murder! I'd need a 300H or so choke to get 20Hz out of it.
I don't need much power, and the B+ can be whatever it needs to be, but I'd like to keep the bias current around 50mA, so larger pentodes are out of the question... I mean, I guess I could run an el34 at 400v and 50mA, and then have the lower 15K Ra and a similar sound...
Anyway, any suggestions?
I don't need much power, and the B+ can be whatever it needs to be, but I'd like to keep the bias current around 50mA, so larger pentodes are out of the question... I mean, I guess I could run an el34 at 400v and 50mA, and then have the lower 15K Ra and a similar sound...
Anyway, any suggestions?
If you use local feedback from the plate of the EL84 to the plate of the driver tube (for an example see RH84 posts) the effective plate resistance drops in proportion to the EL84 transconductance (s) and feedback ratio 👎.
if s = 10,000 and n=0.15, then Rp (effective) = 100/0.15 = 666R.
In Mr Kitic's circuit n is approx 0.1, so Rp(effective) = 1000R
Nice figures for a parafeed choke n'est ce pas?
regards
if s = 10,000 and n=0.15, then Rp (effective) = 100/0.15 = 666R.
In Mr Kitic's circuit n is approx 0.1, so Rp(effective) = 1000R
Nice figures for a parafeed choke n'est ce pas?
regards
Hi aletheian,
if you want to stay with noval socketed tubes, EL86/6CW5 or their 300mA heater variants PL84 (no typo) / 15CW5 may be of interest.
Philips made a commercial SEPP OTL amp with them, the circuit can be found in the Philips EL86 data sheet at frank.pocnet.net, IIRC.
Tom
if you want to stay with noval socketed tubes, EL86/6CW5 or their 300mA heater variants PL84 (no typo) / 15CW5 may be of interest.
Philips made a commercial SEPP OTL amp with them, the circuit can be found in the Philips EL86 data sheet at frank.pocnet.net, IIRC.
Tom
"If you use local feedback from the plate of the EL84 to the plate of the driver tube (for an example see RH84 posts) the effective plate resistance drops in proportion to the EL84 transconductance (s) and feedback ratio 👎.
if s = 10,000 and n=0.15, then Rp (effective) = 100/0.15 = 666R.
In Mr Kitic's circuit n is approx 0.1, so Rp(effective) = 1000R
Nice figures for a parafeed choke n'est ce pas?
regards"
You need to be careful using this approach. By lowering the resistance of the Pentode you up the requirement for the parafeed cap. 4uf caps of good quality are relatively easy to find, but good quality caps at 100uf and 1000V are not.
If you ignore this issue you can expect a peak in the low to mid frequency response, and a fall of in responsew below the peak - not good.
Shoog
if s = 10,000 and n=0.15, then Rp (effective) = 100/0.15 = 666R.
In Mr Kitic's circuit n is approx 0.1, so Rp(effective) = 1000R
Nice figures for a parafeed choke n'est ce pas?
regards"
You need to be careful using this approach. By lowering the resistance of the Pentode you up the requirement for the parafeed cap. 4uf caps of good quality are relatively easy to find, but good quality caps at 100uf and 1000V are not.
If you ignore this issue you can expect a peak in the low to mid frequency response, and a fall of in responsew below the peak - not good.
Shoog
Depending on how you feel about sand an IXYS chip might work well here too. If nothing else it permits playing with the concept for little money.
I have done the chokeless parafeed thing using the IXYS IC's and it works well. There are however a few drawbacks. Since there is no choke the power supply voltage must be raised to supply the headroom. The B+ voltage must be higher than the highest expected positive excursion at the plate of the output tube. For a 2 watt amplifier using a 45 tube I needed 500 volts. This is about the most that you can put through the IXYS chip. The chip will dissipate about 10 watts to do this and it will require a heat sink. The tab on the chip is connected to B+, this makes it almost impossible to insulate it from the heat sink, and you will use a lot of bad language when you touch it!
If you get past all of these drawbacks, it does sound sweet, even with a cheap OPT.
There has been a discussion of using an active load, including some novel CCS concepts proposed in another thread. I am going to experiment with some of these ideas as soon as I have my lab back up.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=67437
If you get past all of these drawbacks, it does sound sweet, even with a cheap OPT.
There has been a discussion of using an active load, including some novel CCS concepts proposed in another thread. I am going to experiment with some of these ideas as soon as I have my lab back up.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=67437
Those are all cool ideas, but before I get all 'fancy,' I think I'll try the 'ol choke plate load first... just to start at the beginning. I may try both the NFB/EL84 thing, and the EL34 thing. I think I have an EL86 around somewhere too, so maybe I'll kick that around to see how it sounds. I have a sinking feeling that I'll be hearing more of the choke quality in the output than the tube's tonality, but I guess I'll see.
Are there any other feedback techniques to lower the output tube's effective internal resistance other than feeding it to the driver? i suppose that would work well, but I was planning on using a 12ax7 or 12at7 there, and that FB worries me. I could easliy use a different driver triode... i just usually experiment with the ones that I can pick up down the street since many meet their untimely demises during the tweaking stage.
How much power do you actually need? If it's really only a couple of watts or less you could run the EL84 in triode connection and Rp under those conditions would be around 2K or perhaps less. Triode connected EL84's sound pretty good imho..
Kevin
Kevin
kevinkr said:
I thought about that. The speakers that I am building are very low sensitivity... and very low power handling too... not a good combination. But they'll take 15w and and have a raw sensitivity of 82dB. Since this is just for kicks, I am not too concerned with volume, but I think that in triode mode, they won't push much air at all.
If you look at Alexander Kitic's RH84 schematic you will see that he uses a 12AT7 driver. A 12AX7 could easily be substituted with a few resistor changes.
I have absolutely no idea why super large parafeed capacitors in the order of 1000uF would be required if you converted this circuit to parafeed.
Rp(effective) = approx 1000R, which is lower than triode mode. Just start around 3uF and experiment for the value that suits you best. Should work nicely.
regards
I have absolutely no idea why super large parafeed capacitors in the order of 1000uF would be required if you converted this circuit to parafeed.
Rp(effective) = approx 1000R, which is lower than triode mode. Just start around 3uF and experiment for the value that suits you best. Should work nicely.
regards
mach1 said:
Hey, well look at that. I think I'll give it a whirl. That feedback is going to lower my driver gain though right? By about 75% just eyeballing it. so I'll have to compensate to get full power from the EL84... man! that is really pushing a 12ax7. I guess I would up the B+ and bump the Rp to 100K or so. I guess the feedback could be dropped a little too by increasing that 100k feedback resistor a bit. i would be comfortable with the EL84's resistance if it were under 10K, so the feedback ratio could be quite a bit lower.
I'll just have to crunch the numbers and see.
Another option suggested by Thorsten is to fix bias the EL84 and use the OPT secondary as the cathode load. It does result in a little DC on the speaker out terminals (though arguable not really worse that an SS amp's electrolytic leakage) and of course sets a limit to the amount of available feedback. Unless your OPT has 70 volt taps. =D On the plus side it doesn't ask any more of the driver (I think).
Haven't tried it yet but will soon with a triode EL84.
Haven't tried it yet but will soon with a triode EL84.
You asked if there were any other techniques to lower the output tubes effective internal resistance. Well the obvious answer is feedback, but the (so far) unmentioned feedback path is to connect the secondary of the output transformer in series with the tubes cathode. This has been done for years, and I have tried it on several SE amps. The results range from useless to amazing, depending on the tube and the transformer.
You seem like a technical person, and there is an excellent technical paper on this technique on Lundahl's web site:
http://www.lundahl.se/pdfs/claus_byrith/cb-amplifier_8wse.pdf
This paper focuses on an EL34 amp but the principle can be applied to any tube. I find it useful for improving the sound of cheap transformers.
You seem like a technical person, and there is an excellent technical paper on this technique on Lundahl's web site:
http://www.lundahl.se/pdfs/claus_byrith/cb-amplifier_8wse.pdf
This paper focuses on an EL34 amp but the principle can be applied to any tube. I find it useful for improving the sound of cheap transformers.
If you are worried about getting DC across the output, wire the cathode resistor to ground and the cathode bypass capacitor to the tx secondary, thereby creating AC feedback only.
regards
regards
"I have absolutely no idea why super large parafeed capacitors in the order of 1000uF would be required if you converted this circuit to parafeed. "
I think I said 100uf at 1000V (depending on the supply voltage). Lowering the output impedence of the tube means you need a larger cap to carry the signal, just as lowering the grid load of a tube increases the needed coupling cap. This is basic theory and applies equally to the output stage. Someone more knowledgeable can correct me if I'am wrong.
Shoog
I think I said 100uf at 1000V (depending on the supply voltage). Lowering the output impedence of the tube means you need a larger cap to carry the signal, just as lowering the grid load of a tube increases the needed coupling cap. This is basic theory and applies equally to the output stage. Someone more knowledgeable can correct me if I'am wrong.
Shoog
A 2A3 has lower plate resistance (750R) than the effective plate resistance (1000R) of the EL84 in the circuit I was describing. Parafeed caps for 2A3 SETs run in the order of 2uF - 7uF, with 3 -4 uF being average.
I have never heard of a parafeed cap in the order of 100uF being used for any application. Have a look at the Valve magazine website for an article on the theoretical derivation of parafeed cap size.
regards
I have never heard of a parafeed cap in the order of 100uF being used for any application. Have a look at the Valve magazine website for an article on the theoretical derivation of parafeed cap size.
regards
tubelab.com said:
It didn't occur to me until afterwards, wouldn't cathode feedback reduce the maximum Pout of an EL84 stage? The stage gain is reduced but the available grid swing, set by the DC differential between cathode and grid, remains unchanged. Less ouput voltage before the grid goes positive, therefore less power out without serious rebiasing. Is that correct? Does an EL84 have enough bias range to compensate?
- Status
- Not open for further replies.