• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

Jens Rasmussen Leach clone group buy

Thanks for the replies. Some good info and good ideas but I still have some questions remaining. As per the SOAR circuit... I've got a good chunk of (re)reading done on the Sloan book I have and studied Leach's write up and I think I will modify this aspect to match Sloan's writings.

As per the DC SOAR curves of the c5200/a1943, they should be able withstand 1A at 70V. With 8 devices, that makes it 8A - into a 4R load, it drops 32V into the load (different if it's highly reactive). It still seems to give ample margin with +/-75V rails - Vce is only 43V at 1A/per device into a 4R load.

Was still wondering about rail voltage, as I'm still considering the 50V transformer to give me ~ +/-70V rails. Seems completely do-able component selection wise but some of the traces on the PCB look pretty close to be using those voltages. Anyone have any insight on this? Second I still haven't figured out if this board requires dual supplies or is designed that is 'can' use dual supplied or supply the front end from the main rails... or requires both. OK well I know it doesn't require both to operate but what is optimal?

There are two traces that go between the 7.5mm pitch pins of the large filter caps. There is more than 20 mils Cu-Cu clearance at all those locations, and the traces are covered by soldermask. You can always spray on a shot of conformal coating at that location if you want more insurance. Clearances at all other locations are significantly more - it can't get better than the pin-to-pin clearance of a TO126 device at any rate (in this case, 2sb649a/2sd669a pre-drivers or the bf470/471 VAS devices).

It requires only +Vs/GND/-Vs, i.e. dual rails. Depending on the value of Vs, two biasing resistors for the zeners/front-end have to have their values adjusted to keep the bias current through the zeners at a nominal 3 mA (?) as per the online notes for the original Leach schematic. I think you'll be fine with just 4 pairs for +/- 70V rails, the additional outputs are mainly for insurance with slightly higher rails (up to +/- 85V is within reach, but I'd stay with +/- 75V even with 8 pairs).

I've seen some late '90s Sherbourn/Emotiva monoblocks with only 4x c5200/a1943, and they're bulletproof even after 15 years with +/- 70V rails. Of course, it depends on the kind of loads they drove during that period.
 
Linuxguru, what I had reservations about is when Vce would exceed 70V (-'ve portion of wave), the leach version just limits current in this region. If the output rail was at -50v, Vce (for c5200) would be 120V. If the C5200 turned on at this point the SOAR says it can only take about 150mA per device which the standard SOAR circuit would not protect the devices. I agree a highly unlikely situation but still possible with a highly reactive load. Less than $2 of additional components could protect against this so I don't see a reason not to investigate it, even if it is likely never to be needed. Could be me being over paranoid as well but thanks for the reality check.

As a side note, I see you mentioned differenct part #'s than what are in the BOM I have. 5200/1942 pair vs. 3281/1302, I didn't realize the latter were renumbered. Probably explains why I have noticed limited availability of the 3281/1302's. Thanks for that. I also see you mentioned the 2SB49A/2SD669A and BF470/471's vs. the MJE340/350 that are in my BOM. Are these replacements as well or just preferred devices? Are the BC556/546 still the preferred devices for the diff. pair?

Thanks for your help.
 
... the standard SOAR circuit would not protect the devices. I agree a highly unlikely situation but still possible with a highly reactive load. Less than $2 of additional components could protect against this so I don't see a reason not to investigate it, even if it is likely never to be needed. Could be me being over paranoid as well but thanks for the reality check.

Jens' current limiting circuit differs from the Leach (v4.5), but I have no idea if it protects against the scenario you mentioned. If you're able to retrofit the Sloan or similar upgraded current limiter, please post the updated circuit here.

As a side note, I see you mentioned differenct part #'s than what are in the BOM I have. 5200/1942 pair vs. 3281/1302, I didn't realize the latter were renumbered. Probably explains why I have noticed limited availability of the 3281/1302's. Thanks for that. I also see you mentioned the 2SB49A/2SD669A and BF470/471's vs. the MJE340/350 that are in my BOM. Are these replacements as well or just preferred devices? Are the BC556/546 still the preferred devices for the diff. pair?

c5200/a1943 are easier to obtain from multiple sources (ST, Fairchild, Toshiba, etc.) - they're the commodity TO-247 of choice, but have worse SOAR compared to some of the On Semi NJW series. The driver is also upgraded to TO-247.

I think MJE340/350 should be fine for the pre-drivers, just check the pinout carefully when mounting them - ECB vs. BCE, etc. The VAS devices are BF470/471 or 870/871 (same die, different package). Again, other TO-126 choices are possible for the VAS, like some of the Sanyo (now OnSemi) Chroma pairs, but verify pinout as above.

For the input pair, BC546B/556B with hFE ~= 300 seems to be the most widely available pair, while the alternatives like MPSA42/A92 seem to have lower hFE. There are numerous Japanese small signal pairs (like c2240/a970) that are arguably better, but with BCE pinout instead of CBE.
 
For the input pair, BC546B/556B with hFE ~= 300 seems to be the most widely available pair, while the alternatives like MPSA42/A92 seem to have lower hFE. There are numerous Japanese small signal pairs (like c2240/a970) that are arguably better, but with BCE pinout instead of CBE.

I can vouch for the c2240/a970 pair. They are very quiet and their very high hFE can really minimize DC offset. They do have a different pinout to consider, though it was absolutely no trouble to arrange the pins to fit.
 
Leach's IV protection is single slope.
Jens' version is a two slope.
The Leach can be converted to the two slope by adding one resistor.

I don't know why Leach missed out that resistor.

IV protection was not Leach's forte. He claimed he never put it in his originial amp, and never had a problem. But he was appreciative of Michael Kawanuka's SOA Protection article I sent him.
 
This might be an old thread but remains and contains extremely invaluable information to the design/improvements/replacements and changes to the late Dr. Leach original Leach amp as well as the Double Barrel amp.
->Leach Amp Plans - Part 1
->Leach SuperAmp Plans

However for future DIY'ers until EDA companies can get their "Shieeyte" straight and stop trying to corner the market on 'drafing/3Dmodel' when the rest of the world CAD file exchangeable formats (IGES, STEP, STL, DXF, etc.) are surpassing and actually making more $$.

Hobbyist be informed.

Attempted to import the Pavouk designed Eagle schematic & Board files into KiCAD in order to increase the Eagle 4" X 4" board limit size.
The Leach amp 200W amplifier

Jens 6 & 10 output pairs final version pdf's and gerber's with drill files are on the waybackmachine website ->http://www.delta-audio.com/index.htm . Jens PDF files of various versions are scattered throughout this forum thread but sadly all Eagle schematic and board files were lost when his hard drive crashed.

At this time KiCAD does not import Eagle schematic or board files.

Having to resort to redrawing and re-layout from scratch.
Forcing everyone to reinvent the wheel is the most stupid thought process any "smart" human/company could possibly conceive.

If anyone has anything better to suggest, I am willing to try.

And, no I will not register to an opensource community just to report a bug within software that has a 'Final' release that they still contend actually imports Eagle .sch & .brd files.

I was using Allied DesignSpark previously, however they have issues with maintaining user accounts when a new version is released and almost impossible to register the update. Been down this road with them since release version 3.6. I can not register this latest release version 8. I am not calling them again for the 5th time to obtain support for their inability to maintain user "registered" accounts.

KiCAD version 5.1.2-2, wxWidgets version 3.0.4, Unicode and Boost 1.68.0.
Eagle version 9.4.2-2.
Win7 Ultimate (build 7601, Service Pack 1 & 3), 64-bit edition.
AMD 8500 Black edition (never OC'd)
Raijuntek 350W Heatsink
32G G.Skill ram (never OC'd)
(2X) Radien RX480 4G video (never OC'd)
1200W RaidMax Modular PSU
(3X) internal 2TB WD harddrives
1 external 4TB WD harddrive
1 external 3TB Seagate
 
Last edited:
LEACH

Leach still going strong;);)
 

Attachments

  • WhatsApp Image 2019-07-22 at 20.33.48.jpeg
    WhatsApp Image 2019-07-22 at 20.33.48.jpeg
    169.5 KB · Views: 256
This might be an old thread but remains and contains extremely invaluable information to the design/improvements/replacements and changes to the late Dr. Leach original Leach amp as well as the Double Barrel amp.
->Leach Amp Plans - Part 1
->Leach SuperAmp Plans

However for future DIY'ers until EDA companies can get their "Shieeyte" straight and stop trying to corner the market on 'drafing/3Dmodel' when the rest of the world CAD file exchangeable formats (IGES, STEP, STL, DXF, etc.) are surpassing and actually making more $$.

Hobbyist be informed.

Attempted to import the Pavouk designed Eagle schematic & Board files into KiCAD in order to increase the Eagle 4" X 4" board limit size.
The Leach amp 200W amplifier

Jens 6 & 10 output pairs final version pdf's and gerber's with drill files are on the waybackmachine website ->http://www.delta-audio.com/index.htm . Jens PDF files of various versions are scattered throughout this forum thread but sadly all Eagle schematic and board files were lost when his hard drive crashed.

At this time KiCAD does not import Eagle schematic or board files.

Having to resort to redrawing and re-layout from scratch.
Forcing everyone to reinvent the wheel is the most stupid thought process any "smart" human/company could possibly conceive.

If anyone has anything better to suggest, I am willing to try.

And, no I will not register to an opensource community just to report a bug within software that has a 'Final' release that they still contend actually imports Eagle .sch & .brd files.

I was using Allied DesignSpark previously, however they have issues with maintaining user accounts when a new version is released and almost impossible to register the update. Been down this road with them since release version 3.6. I can not register this latest release version 8. I am not calling them again for the 5th time to obtain support for their inability to maintain user "registered" accounts.

KiCAD version 5.1.2-2, wxWidgets version 3.0.4, Unicode and Boost 1.68.0.
Eagle version 9.4.2-2.
Win7 Ultimate (build 7601, Service Pack 1 & 3), 64-bit edition.
AMD 8500 Black edition (never OC'd)
Raijuntek 350W Heatsink
32G G.Skill ram (never OC'd)
(2X) Radien RX480 4G video (never OC'd)
1200W RaidMax Modular PSU
(3X) internal 2TB WD harddrives
1 external 4TB WD harddrive
1 external 3TB Seagate

I'm still watching this thread as well, and some years ago I think I mostly re-created in eagle the design that Jens was doing, because he was not sharing the eagle work and I wanted to play with that.

I don't recall which version I did this with, and I may not have it done 100%, but might be pretty close to done, if I could put my hands on it...

I'm still using eagle v6.1, and I never found a stable version of it for mac since. Still, it should be fine and that design should still be usable.

I did my own leach based design as well years ago, including my own mechanical design in CAD. It's slightly different from the original leach design, especially because of the bias sensor diodes, which I don't like the mounting type of... But rather similar anyways.

If there is interest in this leach design, wether it's Jen's or whatever, I could dig back into it and maybe we could cooperate.

I too dislike very much the behavior of those software maintainers that take control of freely available stuff and then take too much control over it and make it far less freely available... I don't like how autodesk has overly taken control over eagle.

I barely tried kicad years ago and didn't really like it, so I stayed with eagle, but didn't upgrade. I'm not paying autodesk anything.
 
Attempted to import the Pavouk designed Eagle schematic & Board files into KiCAD in order to increase the Eagle 4" X 4" board limit size.
The Leach amp 200W amplifier

Jens 6 & 10 output pairs final version pdf's and gerber's with drill files are on the waybackmachine website ->http://www.delta-audio.com/index.htm . Jens PDF files of various versions are scattered throughout this forum thread but sadly all Eagle schematic and board files were lost when his hard drive crashed.
...
If anyone has anything better to suggest, I am willing to try.

I worked with the Pavouk layout - it has some netlist and pinout errors. This is the
2 + 2 transistor monoblock layout with output protection circuitry. I modified many things,
and went with widely available NJW0302/0281 output pairs.

However, the layout is still a .brd only file without a matching schematic, and may still have errors. The parts designations and values haven't been fully filled in.

It will import into Eagle 4.x (which sadly seems to have vanished from the official download sites). From memory the layout is perhaps just within 4" x 6", but I need to dig it out from some legacy system and verify. This a single-sided layout with maybe half-a-dozen jumpers, and may just be within reach of a capable toner-transfer enthusiast.

If you're willing it take it up, I can post or email the .brd file to you with no warranties that it is functional - I think I've fixed all the errors, but it hasn't been fabricated and prototyped yet.
 
I found one old design that is a remake of Jen's leach amp with 5 pairs (on a single board), in eagle 6.1, and it looks pretty much done.
At some point, he made a version that had 2 boards, kindof, with a part of it that can be broken off to make more than one version. I can't recall at the moment how this was, but it had to be at least the 6 pairs version.
Don't know for sure yet if this 5 pair version included the VI limiter, but I'll double check to make sure.
 
Turns out that 5 pairs design version that I found does include the VI limiter and must've been the final version from Jens at the time.

It would probably be best to double check everything to make sure, but this might be pretty much the same as the last iteration of his design.

Making a 6 pairs version shouldn't be much trouble. And I don't think going through the trouble of making a board with breakout parts is worth the effort.

This 5 pairs version I think was aimed at the plastic case for the outputs.
 

Attachments

  • leachamp5p sch.pdf
    66.2 KB · Views: 173
  • leachamp5p layout.pdf
    182.5 KB · Views: 160