labjr said:
You can affect the measured performance by tweaking the 10R and 0.47R resistor values -- but you'd better have a darned good THD% analyzer, and a lot of time and patience.
As I have said before, out of the box, with no tweaking of values the THD% of the amplifier is very low. I matched the 10R resistors with my HP3456A DVM -- matched to within 0.01 ohm -- this did not affect the performance. What I did to optimize values was build a board in which I could parallel in values of R until the THD% was at a minimum.
every time you change one of these values you have to rebias the amplifier and let it come up to a steady temperature.
early in this thread I mentioned that there is an annoying HF peak -- it's just below the a.m. broadcast band so you won't see it on a sound-card analyzer -- this can be nulled out by using separate gate resistors for the JFETs.
you could probably tweak the values of the gate-stoppers 47R -- if you wanted to use the amplifier at RF.
I am going to guess that the difference is attributable to the Fairchild output devices vis a vis the IRF/Vishay devices. My amp also slews a bit faster than the one described in the AX article.
My curiosity was aroused by an application note I saw which indicated that trimming the resistors around a symmetrical pair of MOSFETs could "even them up".
The JFETs I use are Toshiba 2SJ74BL and 2SK170BL -- these were purchased from Newark a couple years ago, while you could stilll get them in quantity.
I can only tell you what I observed. I looked at it with an HP 3577 network analyzer...could also be the little bit of inductance in R11,R12. I've used some Ohmites which are for switching power supplies, I've also looked at Holco, Mills, Dales but the Ohmites (for the price) get the heat off the board.
I am using 4 channels of F5 and am quite happy with how they sound.
But I am still perplexed as to the reason for the discrepancy between Charts 7 and 11 in the AX article. The 2W THD% should be identical ...
I can also tell folks that the sources of the MOSFET's like to see a nice a.c. ground. I use the ELNA Silk 220u/35V
My curiosity was aroused by an application note I saw which indicated that trimming the resistors around a symmetrical pair of MOSFETs could "even them up".
The JFETs I use are Toshiba 2SJ74BL and 2SK170BL -- these were purchased from Newark a couple years ago, while you could stilll get them in quantity.
I can only tell you what I observed. I looked at it with an HP 3577 network analyzer...could also be the little bit of inductance in R11,R12. I've used some Ohmites which are for switching power supplies, I've also looked at Holco, Mills, Dales but the Ohmites (for the price) get the heat off the board.
I am using 4 channels of F5 and am quite happy with how they sound.
But I am still perplexed as to the reason for the discrepancy between Charts 7 and 11 in the AX article. The 2W THD% should be identical ...
I can also tell folks that the sources of the MOSFET's like to see a nice a.c. ground. I use the ELNA Silk 220u/35V
How do you tell if 2SK170BL are real or fake? Reading a few threads now, but no one really seems to know. Do we just go with 'sounds good.. measures good.. price was good' and leave it at that?
I think Futurlec and MCM are good companies but maybe they are fakes since price is quite different from the 2-3 dollars charged elsewhere.
Can you all report how you found out your 2SK170BL was fake if you ever ended up with fakes?
Uriah
I think Futurlec and MCM are good companies but maybe they are fakes since price is quite different from the 2-3 dollars charged elsewhere.
Can you all report how you found out your 2SK170BL was fake if you ever ended up with fakes?
Uriah
http://www.acme-sales.net/acmecart/...p_id=2SK170BL&xm=on&country=USA&ppinc=search2
http://octopart.com/search?q=2sk170
First link is about .69 each then the next are all a few dollars.
Uriah
http://octopart.com/search?q=2sk170
First link is about .69 each then the next are all a few dollars.
Uriah
udailey said:
Measure the transconductance. The next nearest JFET is about 25% of the k170's
K170's and J74's would come packed in a box of 200, and there was a tape sealing the box.
An externally hosted image should be here but it was not working when we last tested it.
Re: PS with bjt question (more PS questions)
Well, it seems that this circuit is just fine as shown. I'm guessing that my problem was that when I mounted the BJT transistors on the heat sink, that one (or more) transistors slipped off their insulation or that there was some debris between the transistor and the sink that cut through the insulator. (I mount all the transistors on the same heat sink.) Basicly, I blew out all of the MOSFETs (the "regulator" and output MOSFETs), and the power supply fuse. After replacing the output FETs, and the "regulator" BJT (and cleaning things very well), the circuit works normally, at least for the last 15 hours or so...
Andrew T is right about the 120R vs 240R resistors. Thanks for the input Andrew!. But in this circuit, I found that the 240R works and with the 5K pot, provides more adjustment.
My unregulated supply is about 28V. The 317/337 dropout is about 2V. The BJT looses about 1V. So I've set the regulated supply to be 25V. This circuit provides less voltage loss than the MOSFET based circuit and I believe it has better regulation than the zener or shunt regulator based references. (I'm not trying to convert anyone, I'm just reporting my results and opinions.) Now it is time to do the other channel.
What I'm most happy about is that my understanding seems correct, and my confidence is restored (to try more new things!).
Thank you all for your support,
Robert
audiorob said:
Well, it seems that this circuit is just fine as shown. I'm guessing that my problem was that when I mounted the BJT transistors on the heat sink, that one (or more) transistors slipped off their insulation or that there was some debris between the transistor and the sink that cut through the insulator. (I mount all the transistors on the same heat sink.) Basicly, I blew out all of the MOSFETs (the "regulator" and output MOSFETs), and the power supply fuse. After replacing the output FETs, and the "regulator" BJT (and cleaning things very well), the circuit works normally, at least for the last 15 hours or so...
Andrew T is right about the 120R vs 240R resistors. Thanks for the input Andrew!. But in this circuit, I found that the 240R works and with the 5K pot, provides more adjustment.
My unregulated supply is about 28V. The 317/337 dropout is about 2V. The BJT looses about 1V. So I've set the regulated supply to be 25V. This circuit provides less voltage loss than the MOSFET based circuit and I believe it has better regulation than the zener or shunt regulator based references. (I'm not trying to convert anyone, I'm just reporting my results and opinions.) Now it is time to do the other channel.
What I'm most happy about is that my understanding seems correct, and my confidence is restored (to try more new things!).
Thank you all for your support,
Robert
you need to subtract half the peak to peak ripple voltage from your 28Vdc supply.
You also need to find how low that supply sags on worst case loading.
You also need to find what the mains supply voltage does to your 28Vdc supply.
A regulated 25Vdc from a 28Vdc supply with all those variables will not work in all operating conditions.
If the 28Vdc is coming from 20Vac then you should be looking at ~20Vdc to 21Vdc as the regulated voltage from that transformer.
If you really want 25Vdc then use a 24Vac or 25Vac transformer.
You also need to find how low that supply sags on worst case loading.
You also need to find what the mains supply voltage does to your 28Vdc supply.
A regulated 25Vdc from a 28Vdc supply with all those variables will not work in all operating conditions.
If the 28Vdc is coming from 20Vac then you should be looking at ~20Vdc to 21Vdc as the regulated voltage from that transformer.
If you really want 25Vdc then use a 24Vac or 25Vac transformer.
Audiorob, why not to try a capacitance multiplier whose drop is a few volts. Would it not be nice for you?
http://sound.westhost.com/project15.htm
http://sound.westhost.com/project15.htm