Hi all,
Fixing an older B&K amplifier here that has a pair of 2SK371s as the input pair. I'm seeing erratic offset voltage, and one of the 371s seems exceptionally thermally sensitive, as in merely touching the case with a finger causes 100s of mV offset which may or may not resolve in a few minutes. I was looking through the Linear Systems offerings and it looks like the dual 2SK389D would work though the transconductance is lower by half, and provide better thermal stability to boot. Anyone suggest a better sub for the 2SK371?
I appreciate the help!
Howie
Fixing an older B&K amplifier here that has a pair of 2SK371s as the input pair. I'm seeing erratic offset voltage, and one of the 371s seems exceptionally thermally sensitive, as in merely touching the case with a finger causes 100s of mV offset which may or may not resolve in a few minutes. I was looking through the Linear Systems offerings and it looks like the dual 2SK389D would work though the transconductance is lower by half, and provide better thermal stability to boot. Anyone suggest a better sub for the 2SK371?
I appreciate the help!
Howie
Hi Howie,
2SK369 is the same part on to92. If it fits then you can try this epray usa seller, part looks genuine.
2SK369 is the same part on to92. If it fits then you can try this epray usa seller, part looks genuine.
Dang!
Thanks so much Chris! I'll let you know if the ones I get from eBay work labeled:
"2SK369BL Original New Toshiba Transistor USA SELLER NOS PART HIGH IDSS"
At least it is a USA seller so I can likely get a refund if they turn out to be bogus. I am also really interested in trying the Linear System K389Ds, I greatly prefer the thermal tracking of dual devices. Looking at the schematic I don't think the substrate diodes will cause any problems in this application. Of course mounting the two K369BLs face to face with thermal compound between them and a zip tie around the two works pretty well too.Cheers!
Howie
Well,
Those 2SK369BLs I bought looked genuine, and definitely old stock if the dried adhesive on the strip is any indication. I installed two that were close in IDSS, put a dab of thermal compound between them and cinched a ziptie around the two. When turned on and offset zeroed, very little drift, so that fixed the offset drift issue. Blowing on the PCB no longer causes hundreds of millivolts of offset.
On the test bench I swept the channel and all looks OK, clean waveform <0.004% THD+N, 220 W out at onset of clipping. Progress!
Cheers,
Howie
Those 2SK369BLs I bought looked genuine, and definitely old stock if the dried adhesive on the strip is any indication. I installed two that were close in IDSS, put a dab of thermal compound between them and cinched a ziptie around the two. When turned on and offset zeroed, very little drift, so that fixed the offset drift issue. Blowing on the PCB no longer causes hundreds of millivolts of offset.
On the test bench I swept the channel and all looks OK, clean waveform <0.004% THD+N, 220 W out at onset of clipping. Progress!
Cheers,
Howie
HELLO...HELLo...hello...is anyone home? Father, it has been four months since my last post and I have sinned. I am restoring an old but beautiful and wonderful sounding if memory serves; Luxman M4000. I made a post in SOLID STATE, no replies yet so I thought I'd bump y'all and repost it here...
I'm working on my M4000 here, and as is almost always the case, schematics, board layouts and the actual equipment have differences. This M4000 is interesting to me that in the the bias circuitry is quite different from the schematic:
.
The SV02 dual diode D401 in the base circuit of Q401 the bias spreader transistor was jumpered out from the factory, and other M4000s I see in YT vids have the SV02 installed and touching the heatsink. I know SV02s are known to be unreliable, so that is potentially a good thing, but only if the bias is thermally stable elsewise.
In this M4000 the heatsink mounted bias spreader Q401 is not heavily bypassed as is the case with many other amps I have worked on. The output board layout and silkscreens and schematic as shown above are calling for a paralleled 0.22uF (C401) and 330uF (C402) cap serving that function, but they were never loaded from the factory. The only cap bypassing the bias spreader transistor is a small 0.22 uF (C216/316) on the driver board.
This would cause quite a bit of AC conduction by the bias spreader, especially at lower frequencies, and that 2SC1904 is only rated at 50 mA. The result would be bias modulation with signal amplitude, of course perhaps that bias modulation was found to be desirable in this amp? Alternatively could the factory have accidentally left off the bypass caps from the output boards?
I have looked on the web to see if any other M4000 output boards have C401 & C402 loaded on the output modules, and the couple of M4000 videos I have found look to have a smaller axial cap than the outline on the silkscreen for C402, but I cannot see the value printed on the cap.
Before I go to the trouble of analyzing this with LTSpice I thought I would ask the brain trust if any of you knew the reason for the different versions of M4000 bias circuits. Any ECOs from Luxman on this issue are lost in the mists of time...
Thanks for any help,
Howie
I'm working on my M4000 here, and as is almost always the case, schematics, board layouts and the actual equipment have differences. This M4000 is interesting to me that in the the bias circuitry is quite different from the schematic:
.
The SV02 dual diode D401 in the base circuit of Q401 the bias spreader transistor was jumpered out from the factory, and other M4000s I see in YT vids have the SV02 installed and touching the heatsink. I know SV02s are known to be unreliable, so that is potentially a good thing, but only if the bias is thermally stable elsewise.
In this M4000 the heatsink mounted bias spreader Q401 is not heavily bypassed as is the case with many other amps I have worked on. The output board layout and silkscreens and schematic as shown above are calling for a paralleled 0.22uF (C401) and 330uF (C402) cap serving that function, but they were never loaded from the factory. The only cap bypassing the bias spreader transistor is a small 0.22 uF (C216/316) on the driver board.
This would cause quite a bit of AC conduction by the bias spreader, especially at lower frequencies, and that 2SC1904 is only rated at 50 mA. The result would be bias modulation with signal amplitude, of course perhaps that bias modulation was found to be desirable in this amp? Alternatively could the factory have accidentally left off the bypass caps from the output boards?
I have looked on the web to see if any other M4000 output boards have C401 & C402 loaded on the output modules, and the couple of M4000 videos I have found look to have a smaller axial cap than the outline on the silkscreen for C402, but I cannot see the value printed on the cap.
Before I go to the trouble of analyzing this with LTSpice I thought I would ask the brain trust if any of you knew the reason for the different versions of M4000 bias circuits. Any ECOs from Luxman on this issue are lost in the mists of time...
Thanks for any help,
Howie
Did you mean to write 22uF instead of 0.22uF? The schematic clip you posted shows 22uF. Actually, it says 22uf 250V, which seems an unnecessarily high voltage rating for a 22uF surely-electrolytic in parallel with the 330uF at 25V. I wonder if the schematic value for C401 was meant to be 0.22uF, which would be something other than an elecrolytic, and the 250V rating wouldn't be so surprising. Does the silkscreen actually say 0.22uF for C401?
It's suspicious that these two are not stuffed. I wonder if some "smart" person looking at this in the middle of production was looking for cost reduction ideas and thought "I bet these two caps don't do much, would it sound good enough if we left them out?"
If a distortion measurementwould be easier than spice, you could do it as-is, and again with these (at least the 330uf) added in, that might suggest what to do.
The TV news just presented a bit about a new service. A place to stay with life force energy!
Based on this
https://www.teslabiohealing.com/
They also did interview a critic.
Interesting to me is no power cord needed, but unit only is rated to work for 3 years.
My conclusion is that the most outrageous claims on this website, now appear to be quite reasonable!
Now back to Howie and reality.
Based on this
https://www.teslabiohealing.com/
They also did interview a critic.
Interesting to me is no power cord needed, but unit only is rated to work for 3 years.
My conclusion is that the most outrageous claims on this website, now appear to be quite reasonable!
Now back to Howie and reality.
This is the kind of website that makes me want to puke. Squeeky-ist of clean 'n polish, shadow people with arms stretched skyward, pretty sunsets, stupid meaningless icons, multiple pop-ups appearing in my face...
If it was of the style of, say, https://www.tinaja.com/ - I might be more inclined to believe...anything they might have to say.
Thanks for the reply Benb!
The schematic is blown up from an incredibly tiny size, and the points mostly got lost. Indeed the 0.22 uF @250 V cap at C401 is a tiny film cap matching the silkscreen outline, and is loaded on the board, but the larger 330 uF@25 V location (12 x 25 mm axial) in parallel is unstuffed from the factory. I guess I am worried if there were that many differences between the schematic and boards whodunnit and why? The 330 uF caps would not make a difference to the DC characteristics of the amp so I am OK powering it up and doing low-power tests. I can scope out the current through the bias spreader to see if it is going to be a problem. Ideally it just floats on the signal with little AC differential across it...we'll see.
Interestingly enough Self says nothing about bypassing the bias spreader in his 'Audio Power Amplifier Design' book, and Cordell merely states "...The usual bypass capacitor should be placed across the bias spreader to ensure stability and maintain low spreader impedance to high frequencies..." I tried entering "Usual Capacitor" in the search bar in Mouser, no results. It seems there will be a HP corner set via the bypass cap and gM multiplier impedance, and it should be below the lowest frequency to be amplified.
The M4000 is also interesting in that it is a triple output stage, but the output bias spreader is across the second stage of the triple. The first stage has it's own thermal diode bias compensation.
I'm going to do more research and get back you youse guys.
Cheers!
Howie
