Welcome! Need support, you got it. Or share your ideas and experiences.
Return to MagneQuest/Peerless Forum
24.6.199.104
I finally got the schematic done for the transconductance amp based on gary Pimms Self Bias CCS. This is based on the idea that a slight modification to the CCS turns in into a very good transonductance AMP. I wanted a true transconductance amp to drive my Lowther speakers. This is a concept made popular recently by Nelson Pass.This implementation is a normal parafeed SE output configuration using BCP-15 and EXO-45. The PS has a PGP-8.1 and RGC-6 choke in a simple LC configuration. There is almost 1V of ripple with this supply but due to the high PSRR of the circuit, it doesn't seem to matter, its completely quiet, even with 100db speakers.
One neat thing about this design is you can adjust the gain by changing R1, this actually chages the Gm of the stage. The 82 ohm shown in the schematic lets me get away without any input stage or input step up transformer.
I did some measurements last night, at 1 watt its 0.1% THD, with an almost ideal harmonic series, 2nd being highest and decresing smoothly from there. The distortion and harmonic serie stayed pretty much the same as I changed the output power over a wide range. There is some PS noise but its 60HZ, not 120, so its not coming from the PS ripple, but probably some coupling from the AC heater.
The frequency response is down -3db at 20HZ, (which is quite expected given the inductance of the plate choke) but its also down 1db at 20KHZ, I'm not sure whats causing this, cpcitance, leakage inductance? Any clues?
Since this is designed to be used as a transconductance amp I also measured the distotion and frquency response at different load impedances. At 20 ohms the increased properly to maintain the current flow through the higher load, this is jut what it was supposed to do. Except at the extemes, the low and high frequency droops increased as the load impedance increased. I understand why this is so at the low end, but again I'm unsure about whats happening at the high end.
These measurements brought to light that the 40H really is not enough, its fine for the 5k, but when it rises due to increasing load impedance, its not enough, the response starts drooping just when I need it to be compensating for the rising impedance of the driver. An 80H - 100H choke seems to be in order here.
Hey Mike, what do you have in the area of a 80 - 100H 60mA or so choke? I was thinking of an EXO-03 for the next version (7-8 watts with RH-PF OPT), but at 48H its not really enough given these measurements.
Sound wise its fantastic, deep, powerful, rich, smooth, extended, high resolution, lush, incredible you are thereness.
BTW I've also built a preamp using these techniques, but its all SS, no tubes, no iron (except the PS) so I can't really talk about it here (I'm not sure WHERE to talk about it) it thinks its a tube, but isn't.
John S.
Follow Ups:
John,
I had read about your transconductance amp in the Tube DIY forum, but just recently I saw that there is a schematic in this forum. Your designs are always innovative and draw my attention, wanting to try them all (for my Fostex speakers).Just a quick question on this amp: what is the current draw, in other words, what current value is the CCS set for?
I decided to try some different parafeed caps on the amps. I had been using the 3.3uf Solens that came with the eXcites, but when I ran Voltseconds equations posted in this forum I came up with 8uf. I also ran Mikes calculations and that came up with 10uf.I had some 8uf 1000V oilers in the parts bin so I decided to try them. The immediate impression was much deeper more powerful bass (out of Lowthers!) but rolled off congested highs. After burning in for three days they sound very differrent, the highs are very open, transparent, detailed etc and the great bass stayed. The overall impression now is significantly better than the 3.3 Solens, and what this did to the bass is almost unbelieveable.
Now I want to try some Mundorf Silvers....
motor run caps. They might really surprise you:)
Hi Russ, I put in the motor runs as parafeed caps in last night, so far they sound more open, less "thick", but a little tizzy on the top. I presume I need to let them burn in for awhile.I had to order some small ones, I'd been using 25-50uf ones in PS duty, but they were way too large for parafeed tasks.
My mercury rectifiers came in today, so the next task is try them out!
The schematic looks cool. I actually have all the parts on hand to wire one up right now!What's the power output of this circuit?
A transconductance amp is one where the output current is proportional to the input voltage, no matter what the impedance of the load. The amp adjusts the output voltage so the current through the load remains constant (for a given input voltage).This circuit is about 2 watts (calculated to be 2.4 watts). It will not work well for driving traditional multiway speakers. The standard crossovers are designed for traditional "voltage" amps. So these types of amps are usually used for single driver speakers such as Lowther or Fostex. You might also want to look at Nelson's article at firstwatt.com fro details of the RC network that many speakers need to work well with this type of amp.
It should also work well with headphones, but I have no idea what the network would be.
As far as the MOSFETs go, don't substitute for the Spertex LND150s these have some special properties that few other transistors have. The IRF820s are not as picky, there are many different MOSFETS that will work here.
Thanks for posting that schematic! --keto
operating as a follower into the cathode of a triode in a grounded grid configuration (but not truly since this is where I inject the feedback) with a FET source follower buffer.So it's almost as evil as yours:) Maybe even worse since I have a solid state voltage regulator too!!
But hey, it sounds nice, it's quiet, and the output impedance is quite low. In short I like it and that should be good enough.....right? Besides, it is only those cable nuts that get really nasty when someone likes something different.
So I went first now tell us about your SS pre...okay? I always like hearing about your stuff because...well...it is not the same old same old same old stuff.
Hi John:you might want to do a serch on the MQ forum for the SJPC 60... from memory this Special Japanese Plate Choke was designed to produce > 100 henries of L while carrying sixty mils of dc plate current.
it's not in production right now... we build them for an overseas customer... but check it out and see if it might meet your needs.
If your current requirements were 40 mils... then the Dowdy plate choke (125H at 40 madc) might be the ticket.
gotta go,
The SJPC-60 was rated at 125H @ 60mAThe SJPC-40 was rated at 175H @ 40mA
They were available for a limited time last summer.
Hi Mike,
the next version uses an EL519 which runs high voltage and high current (but not both at the same time, you still have to stay within the plate dissipation), so could go with a lower primary impedance, which means lower inductance needed, but higher current.I could use a TFA-2004 and a BAC 80 (BTW what's the price on the BAC 80, I don't see it listed anywhere).
Any pinstripe TFA-2004s?
Hi John:we can make a pair of 2004 pinstripes... and depending on the power level needed... we have some 2004 Juniors in various flavors... at good prices.
BAC 80 plate chokes are $150 each.
contact me at acrosound at aol dot com for more information or call me at 215-288-4816 in the evenings.
Hi John:
you wrote:
:::but its all SS, no tubes, no iron (except the PS) so I can't really talk about it here (I'm not sure WHERE to talk about it) it thinks its a tube, but isn't.:::
talk away!!!
This post is made possible by the generous support of people like you and our sponsors: