Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Solid State Talk all about solid state amplification.

Hafler DH-200/220 Mods
Hafler DH-200/220 Mods
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 29th May 2019, 12:03 AM   #1591
808eengr is online now 808eengr
diyAudio Member
 
Join Date: Feb 2019
Quote:
Originally Posted by rsavas View Post
No need for Krispy Kreme. The jfets will be a dual footprint using TO-71-6(socketed) and sot-23-6 on the back, in the 2nd rev, the first rev was only the TO-71, agree, there is not much $ to be saved with the sot-23, thus the choice.
+/-97V, man that is pushing it, using 100V spec'd cap parts, yes it is BBQ season for sure.
maybe time order up your jfets, I am still waiting for mine. I asked trendsetter if they could snail mail them to me, they obliged, since they wanted 2x more money to ship than what they were worth. They only had one lsj689 in stock, but I see that they have re-stocked. I have the contact info if you need it.
Here's some info that might be useful. I ordered a few from them TO-71 in the end of March for a low power amp design. I just tested them. My results:

LSK489: 3.41mA (A), 5.46mA (B), 6.04mA (B), 7.80mA (B)
LSJ689: 6.47mA (A), 7.10mA (B), 11.76mA (B-C), 14.56mA(C)

Not the results I was hoping for, but I have two usable sets from purchasing four sets. Too bad they are not available for purchase in different grades like the other Linear Systems fets. Good luck with the project.

info on test setup
Idss tests of LSK489 and LSJ689

Last edited by 808eengr; 29th May 2019 at 12:06 AM. Reason: to-71
  Reply With Quote
Old 29th May 2019, 03:16 AM   #1592
rsavas is offline rsavas  Canada
diyAudio Member
 
Join Date: Aug 2012
Location: Ontario
Need to ask Bob C. if there is any requirement for matching them in the DH-220C design.
  Reply With Quote
Old 1st June 2019, 12:13 AM   #1593
Bob Cordell is offline Bob Cordell  United States
diyAudio Member
 
Join Date: Sep 2006
Quote:
Originally Posted by 808eengr View Post
Here's some info that might be useful. I ordered a few from them TO-71 in the end of March for a low power amp design. I just tested them. My results:

LSK489: 3.41mA (A), 5.46mA (B), 6.04mA (B), 7.80mA (B)
LSJ689: 6.47mA (A), 7.10mA (B), 11.76mA (B-C), 14.56mA(C)

Not the results I was hoping for, but I have two usable sets from purchasing four sets. Too bad they are not available for purchase in different grades like the other Linear Systems fets. Good luck with the project.

info on test setup
Idss tests of LSK489 and LSJ689
When implementing full complementary LTP differential input stages with N-type and P-type dual monolithic JFETs, matching of Idss between the P and N devices is not as important as one might think. In use, the operating Vgs of the P and N devices will go to whatever voltage it takes to cause the device pair to operate at its available tail current.

What is more important in such an input stage is that the transconductance of each LTP be about the same. For devices of a given process, the operating transconductance is mainly a function of operating current, rather than Idss. In other words, two JFETs from the same process may have different Idss and threshold voltage, but the same transconductance. This is because the transconductance parameter Beta in the SPICE model is similar for devices from the same process.

However, it turns out that the transconductance for a P-channel device and an N-channel device may have different Beta, and consequently different transconductance at a given current value. What this means is that even if you use P and N devices with exactly the same Idss, you will not likely have the same transconductance in the P and N pairs, even if the tail currents are the same.

Sometimes, source degeneration resistors will be added to one of the pairs to make the transconductance of the P and N pairs the same at the same operating current.

The observation that gm for devices from the same process depends mainly on operating current rather than Idss makes the paralleling of pairs to reduce noise more practical without resort to close matching, as long as each of the paralleled pairs has its own tail current source. This approach was used to advantage in the VinylTrak MC preamp where four LSK389 devices were connected in parallel to achieve low noise.

That all having been said, I do wish that the LSK489 and LSJ689 parts were made available in Idss-graded versions.

Cheers,
Bob
  Reply With Quote
Old 4th June 2019, 03:01 PM   #1594
AMRoberts is offline AMRoberts  United States
diyAudio Member
 
Join Date: Oct 2007
Default Dimensions for the power supply caps?

Hi,

Does anyone have the dimensions for the stock power supply filter caps (C403 & C404) used in the DH-220? Or if there is already a post in this thread please point me there (my search efforts aren't finding it).

I thought I took measurements the last time I had the cover off to replace the power switch, but that piece of paper didn't make it back into the file with my Hafler kit building instructions :-(.

I'd like to do a "drop in" replacement of the originals, so physically I need to match the diameter, be less-than-or-equal-to the height, and have the same screw terminals, correct? If anybody has the dimensions I can order parts before I pull the amp out of the equipment shelving and take the cover off again.

Thanks,
Alan
  Reply With Quote
Old 5th June 2019, 03:46 AM   #1595
808eengr is online now 808eengr
diyAudio Member
 
Join Date: Feb 2019
Question A few questions about use of LSJ689/LSK489

Quote:
Originally Posted by Bob Cordell View Post
What is more important in such an input stage is that the transconductance of each LTP be about the same. For devices of a given process, the operating transconductance is mainly a function of operating current, rather than Idss. In other words, two JFETs from the same process may have different Idss and threshold voltage, but the same transconductance. This is because the transconductance parameter Beta in the SPICE model is similar for devices from the same process.

However, it turns out that the transconductance for a P-channel device and an N-channel device may have different Beta, and consequently different transconductance at a given current value. What this means is that even if you use P and N devices with exactly the same Idss, you will not likely have the same transconductance in the P and N pairs, even if the tail currents are the same.

Sometimes, source degeneration resistors will be added to one of the pairs to make the transconductance of the P and N pairs the same at the same operating current.

The observation that gm for devices from the same process depends mainly on operating current rather than Idss makes the paralleling of pairs to reduce noise more practical without resort to close matching, as long as each of the paralleled pairs has its own tail current source. This approach was used to advantage in the VinylTrak MC preamp where four LSK389 devices were connected in parallel to achieve low noise.

That all having been said, I do wish that the LSK489 and LSJ689 parts were made available in Idss-graded versions.

Cheers,
Bob
Thanks for the clarification Mr Cordell. The application notes you've written and your book are my most valuable resources. As an electrical engineering student I feel like the art of analog circuitry design and understanding is being lost to Arduinos and breakout boards. I'm struggling to find a professor to work with for an audio based junior/senior design project... if I said the word 'drones' it would not be so tough.

With the explanation provided this makes the floating style of the complementary sound like even more of a challenge and I have a few questions.

Does different output impedance of the LTPs due to degeneration resistors have an effect on the rest of the circuit?

If you do you use independent CCS for the LTPs (or were to try and match gm for a given trail current) is it best to measure amplitude at the LTP or at the VAS? Could using different grade Hfe transistors in the VAS be an alternative way of reaching a balanced push pull voltage before the output stage?

Thanks for your time
  Reply With Quote
Old 6th June 2019, 05:24 PM   #1596
Bob Cordell is offline Bob Cordell  United States
diyAudio Member
 
Join Date: Sep 2006
Quote:
Originally Posted by 808eengr View Post
Thanks for the clarification Mr Cordell. The application notes you've written and your book are my most valuable resources. As an electrical engineering student I feel like the art of analog circuitry design and understanding is being lost to Arduinos and breakout boards. I'm struggling to find a professor to work with for an audio based junior/senior design project... if I said the word 'drones' it would not be so tough.

With the explanation provided this makes the floating style of the complementary sound like even more of a challenge and I have a few questions.

Does different output impedance of the LTPs due to degeneration resistors have an effect on the rest of the circuit?

If you do you use independent CCS for the LTPs (or were to try and match gm for a given trail current) is it best to measure amplitude at the LTP or at the VAS? Could using different grade Hfe transistors in the VAS be an alternative way of reaching a balanced push pull voltage before the output stage?

Thanks for your time
Thanks for the kind words. I am also disappointed that the teaching of analog circuit design seems to be on the decline compared to that for other areas. But there is so much in electrical engineering these days that it is probably a challenge for educators to choose their priorities. Quite a few years back the late Marshall Leach taught an audio-centric undergraduate course at Georgia Tech that was wildly successful. It was based on his book. One of the key things to the popularity and effectiveness of the course was that audio was the hook that got students interested. The material also gave students good analog EE intuition and hands-on experience. Marshall was a great educator. I'd love to see my book get used in some early-level EE courses in 2-year or 4 year colleges. It is obviously not a high-level textbook, but if it got more students interested in analog design it would have served its purpose. I think it could also serve as the basis of some useful lab experiments. We'll see :-).

While the use of degeneration resistors will have some influence on the output of the LTPs, it may not make much difference in comparison to everything else. For example the output impedance of the un-degenerated pair of devices may be higher to start with as a result of their lower operating gm (which was the reason we added degeneration to the other side in the first place). The natural output impedance of P and N-channel devices operating at the same drain current may not be the same anyway. Finally, in typical Miller-compensated amplifier designs, differences in LTP output impedance will tend to matter largely only at lower frequencies below the open-loop bandwidth frequency, where there is less concern anyway. This is largely because the shunt feedback from the Miller compensation capacitor effectively makes the output node of the LTP a low-impedance node anyway.

Definitively don't try to improve matching of LTP gain by fooling around with beta of the VAS transistors. More beta in the VAS transistors, even if one is 150 and the other is 300, is usually better. And, as usual, I recommend using a 2T VAS.

Hope this helps.

Cheers,
Bob
  Reply With Quote
Old 7th June 2019, 01:18 AM   #1597
phase is offline phase  United States
diyAudio Member
 
Join Date: Oct 2004
I think that would be great to have your book as a textbook!
  Reply With Quote
Old 15th June 2019, 05:23 AM   #1598
Rinman77 is offline Rinman77
diyAudio Member
 
Join Date: Mar 2019
Default Questions about Hafler DH500 power supply mods

Hi All,
I just purchased an HF500 off Ebay. When I opened it up I saw that the power supply has been modified; instead of the single large transformer there are two large torroids stacked, (Talema 2185 (0500-1-130)). And instead of the stock capacitors there are 4 x "Capacitor Technology" 3800 uF caps.
Anyway I have a few initial questions: