Those don't look to be cheaper.
TechDIY's quartet of matched pairs at $64 looks like a very good deal. That would be 2 complete output stages, if they are 2 pairs each (is more actually needed?).
They have been matched, which means you can get away with a very low (or none, if you are feeling Nagy 😀 ) source resistor.
TechDIY's quartet of matched pairs at $64 looks like a very good deal. That would be 2 complete output stages, if they are 2 pairs each (is more actually needed?).
They have been matched, which means you can get away with a very low (or none, if you are feeling Nagy 😀 ) source resistor.
Lateral Mosfets
Magnatec/Semelab also makes the 2SK135 and2SJ50 available through RS at a reasonable price. Are these any good?
😕 ckt
Magnatec/Semelab also makes the 2SK135 and2SJ50 available through RS at a reasonable price. Are these any good?
😕 ckt
I should have most of what I need soon, I've ordered from Mouser. I won't be able to afford 64$ worth of outputs, but I can buy 1 or 2 pairs.
- keantoken
- keantoken
A man in a BROWN uniform
off of a BROWN truck
Just brought me a BROWN box
with baggies of about 300 trannies in it.
From Mouser to the land of the house of the mouse.
Are you ready for some FootBAll
I am very disappointed in the results of the CX by OS, but some ground plane and time to investigate the real deal is here. I have two more transformers coming from Newark direct from Hammond. The front end power supply is about done and I am working on make it a dual reg tracking one.
Well the rubber is about to get introduced to the asphalt.
off of a BROWN truck
Just brought me a BROWN box


From Mouser to the land of the house of the mouse.
Are you ready for some FootBAll

I am very disappointed in the results of the CX by OS, but some ground plane and time to investigate the real deal is here. I have two more transformers coming from Newark direct from Hammond. The front end power supply is about done and I am working on make it a dual reg tracking one.
Well the rubber is about to get introduced to the asphalt.

Well there will not be ANY LSK389 A fet's until the end of January according to Linear Systems...😡
C's are out there but not the A's
C's are out there but not the A's

I shouldfford 64$ worth of outputs, but I can buy 1 or 2 pairs.
- keantoken
Hey Keantoken:
Have you had a chance to put any of those mouser parts together?
Just where is the project at this juncture?
I have all the transistors except the outputs, which should be arriving on a business day. I also received a hefty toroid from Ed Lafontaine which should work, though it may be tricky to elevate the upper rails.
- keantoken
- keantoken
Magnatec/Semelab also makes the 2SK135 and2SJ50 available through RS at a reasonable price.
Equivalent numbers, not the exact same but easy and affordable.
Input FET
I have both the duals 2SK146 and 2SK389. Which one would be the better choice as the input pair?
😕
ckt
I have both the duals 2SK146 and 2SK389. Which one would be the better choice as the input pair?
😕
ckt
Has anyone done a thermal analysis of the output section of this amp? What is the thermal resistance of the heat sink? What will the junction temperature be at full power?
I'm guessing that will depend on what you intend to use it for. For normal listening full power isn't necessary. But then again, who would need 80V rails for normal listening... Hmm...
Can't be any worse that the Goldmund, as long as we stick with 4 outputs. If you use less outputs, should probably decrease rails.
Goldmund's logic is starting to confuse even me now though. At 80V you can hardly get 2A out of those laterals at 20Hz, why even consider it? Lower rails to 40V, and you can have over 5A out of them, MUCH more efficient. So why buy 4 pairs for $64, when you could have more power, less heat, and better frontend transistors?
Anyway Sawreyrw, from a technical POV, this amp isn't a shining beacon of indestructable design. As I understand it, potential builders don't intend to do more than treat it lightly. But I think I should have some second opinions on this.
Guys, why have 80V rails when we can halve the rails and have a cooler amp with less risk of destruction? 80V into 8ohms is 800W, those Hitachis can't possibly drive that. The most they can do with 80V rails is 32W (2A into 8ohms) per pair of outputs. Halve rails and you have nearly 200W (5A, 8ohm) with just one pair of outputs.
Does my math make sense?
- keantoken
Can't be any worse that the Goldmund, as long as we stick with 4 outputs. If you use less outputs, should probably decrease rails.
Goldmund's logic is starting to confuse even me now though. At 80V you can hardly get 2A out of those laterals at 20Hz, why even consider it? Lower rails to 40V, and you can have over 5A out of them, MUCH more efficient. So why buy 4 pairs for $64, when you could have more power, less heat, and better frontend transistors?
Anyway Sawreyrw, from a technical POV, this amp isn't a shining beacon of indestructable design. As I understand it, potential builders don't intend to do more than treat it lightly. But I think I should have some second opinions on this.
Guys, why have 80V rails when we can halve the rails and have a cooler amp with less risk of destruction? 80V into 8ohms is 800W, those Hitachis can't possibly drive that. The most they can do with 80V rails is 32W (2A into 8ohms) per pair of outputs. Halve rails and you have nearly 200W (5A, 8ohm) with just one pair of outputs.
Does my math make sense?
- keantoken
keantoken,
It sounds like there is more work to be done in this area. What is the "specified" power output at what load resistances - 4 ohms and 8 ohms? This then will be used to determine the supply voltage and power dissipation for the output section. Finally you can do the thermal analysis and select the heat sink.
In any case, I don't think you should use a suppply voltage way higher than necessary. And you shouldn't design something that can be overloaded and damaged by increasing the power into the rated load.
Rick
It sounds like there is more work to be done in this area. What is the "specified" power output at what load resistances - 4 ohms and 8 ohms? This then will be used to determine the supply voltage and power dissipation for the output section. Finally you can do the thermal analysis and select the heat sink.
In any case, I don't think you should use a suppply voltage way higher than necessary. And you shouldn't design something that can be overloaded and damaged by increasing the power into the rated load.
Rick
Goldmund's logic is starting to confuse even me now though. At 80V you can hardly get 2A out of those laterals at 20Hz, why even consider it? Lower rails to 40V, and you can have over 5A out of them, MUCH more efficient. So why buy 4 pairs for $64, when you could have more power, less heat, and better frontend transistors?
80V into 8ohms is 800W, those Hitachis can't possibly drive that. The most they can do with 80V rails is 32W (2A into 8ohms) per pair of outputs. Halve rails and you have nearly 200W (5A, 8ohm) with just one pair of outputs.
Does my math make sense?
- keantoken
As a clue, back in 1985 or so, i did build a symetrical differential amp using
three pairs of 2SJ48/2SK133 which are the 120V lower grade of the 2SJ50/
2SK135 laterals.
Power supply was +-60V, and it happens that main voltage was increased
in France from 220 to 230V, increasing the PS to about +-63V, and i often
measured a few volt more at iddle...
Transformer was a beefy 620VA toroid along with 4X10000uF.
This amp has to endure the baddest treatements, full power on live stages,
often loaded with 4R , with ambiant as much as 40°C.
yet, it did always work, even it didn t have any current limiting circuit....
Attachments
Goldmund's logic is starting to confuse even me now though. At 80V you can hardly get 2A out of those laterals at 20Hz, why even consider it?
Featherweight amps for featherweight loudspeakers.
Aah, the trusted SK53 again.
Sugden's P180 uses 2pair of Laterals running on near +-80Vdc. The PSU output depends on the bias setting and the mains voltage.I'm guessing that will depend on what you intend to use it for. For normal listening full power isn't necessary. But then again, who would need 80V rails for normal listening... Hmm...
Works well into 8ohms speakers for domestic listening.
Also rated for 4ohm speakers but when tested into 4r0 the PSU falls to ~ +-56Vdc.
I think Sugden deliberately used small smoothing caps and high regulation transformers to protect the amplifier from the effects of abuse.
The Laterals are also very robust.
I doubled the output stage to 4pair and doubled the smoothing capacitance. Sounds and measures better into 8ohms/8r0. I don't have 4ohm speakers but it can now drive 4r0 quite a bit better.
Last edited:
The current version is supposed to be the "high-power" version. For 60V rails the Hitachis should be enough but for 80V I would definitely choose something else.
The idea was that I would develop the high-voltage version first and then downgrade, but using 80V rails for these devices looks insane to me (either no one pointed it out or I seriously lost the memo). That's a good thing in one way though, because my test transformer is 54V.
At 60V the outputs can drive 3A. This means 78W into 8R. At 50V it's 4A, 128W. So my 54V toroid actually seems perfect. Use two pairs for 4R load, 1 pair for 8R (2 pairs default for good measure).
- keantoken
The idea was that I would develop the high-voltage version first and then downgrade, but using 80V rails for these devices looks insane to me (either no one pointed it out or I seriously lost the memo). That's a good thing in one way though, because my test transformer is 54V.
At 60V the outputs can drive 3A. This means 78W into 8R. At 50V it's 4A, 128W. So my 54V toroid actually seems perfect. Use two pairs for 4R load, 1 pair for 8R (2 pairs default for good measure).
- keantoken
Would it be possible to put together a FULL and complete schematic of the ENTIRE amplifier that you are going to build?
Shunts, power supply, TMC, warts, front-end ops etc etc...
I have gotten more than a little lost of the whole she-bang
Shunts, power supply, TMC, warts, front-end ops etc etc...
I have gotten more than a little lost of the whole she-bang
At 60V the outputs can drive 3A. This means 78W into 8R. At 50V it's 4A, 128W. So my 54V toroid actually seems perfect. Use two pairs for 4R load, 1 pair for 8R (2 pairs default for good measure).
- keantoken
Where/how did you get these numbers for volts and amps? (BTW, 3^2*8=72, not 78). Obviously, you haven't done the thermal analysis? Oh BTW, what is the spec on the power output? Who was that guy that kept harping about specs?
I try my best, but what you see is the result of me not having any formal education. I can crunch numbers and simulate, but obviously I'm lacking in how it all comes together in a finished design. (of course it doesn't help when people think the whole project should be done in a week, and to boot it has to be the best amp in 20 years)
Obviously, I can't do this by myself as I have been saying all along.
I figure power specs should follow economy. There is no standard load, so there can't be a standard rail voltage that will suit all loads. Ideally the builder would figure out how many outputs and what rails to use for their specific load. I like the idea of teaching a man to fish. This is DIY, so trying to make a one-size-fits-all would be missing the point perhaps?
BTW, I could do a junction temperature simulation if I knew the die and case thermal capacity, die-case thermal resistance, and case-heatsink thermal resistance. I might be able to derive some of these from the datsheet.
- keantoken
Obviously, I can't do this by myself as I have been saying all along.
I figure power specs should follow economy. There is no standard load, so there can't be a standard rail voltage that will suit all loads. Ideally the builder would figure out how many outputs and what rails to use for their specific load. I like the idea of teaching a man to fish. This is DIY, so trying to make a one-size-fits-all would be missing the point perhaps?
BTW, I could do a junction temperature simulation if I knew the die and case thermal capacity, die-case thermal resistance, and case-heatsink thermal resistance. I might be able to derive some of these from the datsheet.
- keantoken
- Home
- Amplifiers
- Solid State
- Goldmund Mods, Improvements, Stability