The mosfet follies stuff looks alot like the headphone amp I am doing now... only major diffirence is the use of a constant current source on the anode.
This is my 4th version now, useing IRF610, waiting now for RS to phone, so I can fetch more, as I blew one.
I don't know if its called speed, but heck these babies kick some serious but in the drums category, appart from the treble shyness, I think was caused by the cheap 6dj8 I was testing with, I have the Amperex PQ E188CC waiting on my desk with me to see if this clears things up... just actauly noticed I had one in my last valve buffer... lol, I thought it was an ecc88 at the time I build it...
It is basicaly a reworked Sijosae MHHA, but bigger for better heatsinking and less heat dissipated straight onto components and including a socket for you snobs...
. I suspect there will be one more version before final release, as I have no decoupling caps in, and made one or two other small compromises to keep things as cheap and easy as possible..
This is my 4th version now, useing IRF610, waiting now for RS to phone, so I can fetch more, as I blew one.
I don't know if its called speed, but heck these babies kick some serious but in the drums category, appart from the treble shyness, I think was caused by the cheap 6dj8 I was testing with, I have the Amperex PQ E188CC waiting on my desk with me to see if this clears things up... just actauly noticed I had one in my last valve buffer... lol, I thought it was an ecc88 at the time I build it...
It is basicaly a reworked Sijosae MHHA, but bigger for better heatsinking and less heat dissipated straight onto components and including a socket for you snobs...
. I suspect there will be one more version before final release, as I have no decoupling caps in, and made one or two other small compromises to keep things as cheap and easy as possible..
SY said:
thats the kind of information I almost wish I didnt know... heck I've been called a luddite before...
Another complication is that I want to try actively controlling the output stage, the CCS under a cathode follower gives me a control input. VERY similar to the modulated B+ idea that you laid out in the unity gain stage. Possibly crossbread with a bias servo that's capable of class B operation. - I'm a little removed fron actually having a schematic yet though
Now apreciate that this is probably simpler with a Mosfet driver but thats just a little too far from the comfort zone.
Andy
Subminis are readily available on eBay, New Sensor has a good selection and in your neck of the woods Russian milspec surplus is also readily available.. (And mine too actually)
They have very high transconductance which in some applications will translate into lower noise, they also make better followers for the same reason, and are pretty rugged. Some types are more or less microphonic than others..
Good ones to look at are:
6021 (dual triode, low mu of 35, gm of 5400uS, Note low max vp)
5744 (single triode, high mu of 77, gm of 4000uS)
5703 (single triode, low mu of 25, gm of 5000uS)
None of these are particularly expensive or hard to find. All are worth looking at because of their transconductances and relatively high allowable plate currents.
They have very high transconductance which in some applications will translate into lower noise, they also make better followers for the same reason, and are pretty rugged. Some types are more or less microphonic than others..
Good ones to look at are:
6021 (dual triode, low mu of 35, gm of 5400uS, Note low max vp)
5744 (single triode, high mu of 77, gm of 4000uS)
5703 (single triode, low mu of 25, gm of 5000uS)
None of these are particularly expensive or hard to find. All are worth looking at because of their transconductances and relatively high allowable plate currents.
I have come to similar conclusions. I still get email saying that tube amplifiers should not be contaminated with sand. So I set out to build a vacuum tube based PowerDrive. I have built a few of them and none have worked as well as the SS version. Someone asked me for FFT measurements on the PowerDrive circuit alone, so I thought I would compare the SS version and the vacuum tube version. I haven't done that yet because I am still not happy with the vacuum tube version, and I am still testing mosfets for the PowerDrive since my old favorite the 2SK2700 has been banned from the planet (it is not available in an unleaded version).
The best vacuum tube PowerDrive so far uses a 6BL8 for the driver tube. The triode section is the voltage amplifier and the pentode section is the CCS. The follower is currently a pentode wired cathode follower using a 6CW5. The mosfet version still sounds better. No FFT measurements yet.
kevinkr said:
6021 had slipped past me, I've been trying to get my mitts on as many decent ones as I can, they just made the list...
I also got quite fond of the 5977, Low Mu of 16, might be nice to use in a pre-amp of some sort. [Project #276 and counting...]
There are others but I haven't got enough yet
I do see your point though Nordic, with sub-mini's it's sometimes a question of seeing what you can get. If you want your amp to be widely adopted then specifying tubes that are hard to find won't help. Lest we fall into the 6528 trap...
Andy
thought you may enjoy this
An externally hosted image should be here but it was not working when we last tested it.
tubelab.com said:
this is the bain of my life at the moment, for a day job I'm involved in getting assemblies from China into Europe. Not electronic assemblies mind, just pressed steel bits... I dont really follow the argument about 'contaminated with sand' though theres too good an argument for Solid state rectification / CCS's and so on....
Anyway back to the thing that started this thread, What about an EL37? I managed to get a working Mullard valve tester with manual and so on, it has one of these in it as the pass element in the power supply. Good for audio too?
oh and makes the bally 6528 look cheap! But this is probably the place for esoteric tubes.
Andy
There are always people who are highly resistant to change. When I first published the PowerDrive circuit I got a lot of negative email. My favorite was the guy saying I should change my web sites name to Transistorlab. Email was initially about 75% negative. Now two years later after a few early adopters raved about the sound, the email is mostly positive or inquisitive. I get very few negative email about PowerDrive. I have a few more SS/tube hybrid circuits that will eventually make their way on to my web site, after they are thoroughly tested.
There have been some concerns raised about the fact that all silicon devices have a voltage varying capacitance property, which could cause some phase distortion when placed in the signal path. These concerns are valid, and the effects are hard to measure with conventional equipment. For these reasons, choose mosfets carefully based on the capacitances and how they vary with voltage. Find a fet that remains fairly constant in your operating range.
The chart shown in a previous post categorizes some CCS circuits based only on their current VS voltage performance. The LM317 is shown as excellent. Most LM317's loose their effectiveness with increasing signal frequency. They work quite well in cathode circuits where they are often bypassed. I wouldn't use one in the signal path.
SY said:
Mines broken then
'cos it sounds awful. Thing is it SHOULD & so well. It occurred later to try it in place of an EL34 and see if I could "hear" a problem. Well it didnt sound right...
Heck, I'll see if I can get a pair for less money than the average house, if nothing else they'll work in the SE amp..
You know these things have little Ceramic ferrules to space the rods out on the mica. Beautiful!
Andy
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.