MC2120 Driver subs. Has anyone tried TO-220 with success?

[Skywatcher]

Since the TO-3 "modern" subs are NLA, and were "only" a 4A rating, why would a modern TO-220 not work? The 30MHz capability is there with many TO-220 along with higher current. Being on the main heat sink appears to be more for direct thermal tracking and not so much current/wattage. Mounting to the heat sink would not really be an issue, and soldering wiring to the leads relatively easy.
I cannot find the datasheet for the "original", but the Motorola/ON that were used as subs datasheets are available.
Is it possible the MJE15032/33 would function correctly? And with which output transistors?
Or other TO-220 with enough "guts"?
I have 21193/94, 95/96, and 003/4 TO-3 outputs on hand.
Yes, I asked before, but never got any definite answers specific to the MC2120.

 

[Classic_CJ]

There seem to be a few (surplus) vendors out there for the MJ15018/19/20/21 series of original OnSemi 4A devices when I checked Octopart a short while ago, but you have to buy a whole big bunch of them. That being that, has anyone checked Mac factory recently to see if they have any in stock? The last Mac #s I have are 132-246 & 132-247

There's always NTE, where they cross the NTE87 and NTE88 for the above.

https://www.nteinc.com/specs/10to99/pdf/nte87.pdf

 

[Skywatcher]

There seem to be a few (surplus) vendors out there for the MJ15018/19/20/21 series of original OnSemi 4A devices when I checked Octopart a short while ago, but you have to buy a whole big bunch of them. That being that, has anyone checked Mac factory recently to see if they have any in stock? The last Mac #s I have are 132-246 & 132-247

There's always NTE, where they cross the NTE87 and NTE88 for the above.

https://www.nteinc.com/specs/10to99/pdf/nte87.pdf

Those NTE are only 4MHz. Need at least 20MHz, Yes a friend of mine checked with Mac and nada. I saw them on Octopart too, but I don't need that many.
Tried what was once a "reliable" source on fleabay, but now peddle ChinoCrap. I found out they have one address on fleabay, but ship from Bordertown, NJ under another seller name, and it's the same address as yet a 3rd fleabay seller under a different name. Semitronus and activeparts are now confirmed to be one in the same. When I sent them back (bought from both sellers to check them) this was the response... It was the same response from both...

Hello,
Thank you for reaching out. I'm sorry to hear you cannot use these parts. These parts were ordered from a trusted US supplier many years ago and are original parts to my knowledge.

One set had the Motorola logo, the other ON logo and were identical otherwise with wrong format and font. Neither matched any of my older known to be genuine stamping of both M and ON..

So in light of it all is why I am leaning towards a known reliable, genuine TO-220 substitution given the specs.
I wish I owned this unit. If I did, I would experiment with it without hesitation.

 

[JoseHH]

Based on the datasheets for the MJ15020/21, I see no reason why the MJE15032/33 would not work there. Furthermore, to me, it sems strange to use a 4A driver in an amplifier, unless you are driving multiple output transistors. I dont know if this is really the case for this model.
If the owner agrees, the only way to get his amplifier back to working order may be to try the MJE15032/33 as drivers. Are there any other unobtanium parts in this amplifier that could be temporarily substituted to run a test with the ON TO-220 drivers? This can prevent loosing other hard to find parts.

 

[Skywatcher]

Based on the datasheets for the MJ15020/21, I see no reason why the MJE15032/33 would not work there. Furthermore, to me, it sems strange to use a 4A driver in an amplifier, unless you are driving multiple output transistors. I dont know if this is really the case for this model.
If the owner agrees, the only way to get his amplifier back to working order may be to try the MJE15032/33 as drivers. Are there any other unobtanium parts in this amplifier that could be temporarily substituted to run a test with the ON TO-220 drivers? This can prevent loosing other hard to find parts.

This has 4 outputs per channel, 2NPN, 2 PNP. I don't believe there are any "obscure" components on the amp cards. Only one channel is popped so the other won't need anything. Here is the service manual if it helps..

http://www.berners.ch/McIntosh/Downloads/MC2120_ser_b.pdf

 

[JoseHH]

I have never worked on one of these, so I compare it with the sansui AU-D907 (same as AU-919) which also has two complementary power transistors driven by a 1.5A driver (2SD382/2SB537) but with higher voltage rails (+/-56V). I see the schematics now and I understand your hessitation. This McIntosh uses +/-42 Volt rails, so to get 120W (comparable to the AU-919) it will have to pass about 35% more current than the AU-919. As a guideline, scaling up the 1.5A drivers in the AU-919 by 35% yields a little over 2A. The SOA for the MJ15020 shows almost 4A capability at 40V, while the MJE15032 is down to 2A at that voltage, unfortunately it is barely there.
EDIT:
The estimate avobe is a little sketchy.

For 120W and 40V would be an RMS current of 3A in the combined output transistor set. If the outputs were to have hfe=3, they would demand 1 amp of base current which should be within the SOA of the MJE15032. Most likely the gain of the output transistors is going to be in the 10-50 which lowers the drive current needed. So my guess is that the MJE15032/3 should be fine if you can not get genuine NOS MJ15020/21 or MJ15018/9. I searched for other possibilities but found nothing else that could serve.

 

[Skywatcher]

I have never worked on one of these, so I compare it with the sansui AU-D907 (same as AU-919) which also has two complementary power transistors driven by a 1.5A driver (2SD382/2SB537) but with higher voltage rails (+/-56V). I see the schematics now and I understand your hessitation. This McIntosh uses +/-42 Volt rails, so to get 120W (comparable to the AU-919) it will have to pass about 35% more current than the AU-919. As a guideline, scaling up the 1.5A drivers in the AU-919 by 35% yields a little over 2A. The SOA for the MJ15020 shows almost 4A capability at 40V, while the MJE15032 is down to 2A at that voltage, unfortunately it is barely there.
EDIT:
The estimate avobe is a little sketchy.

For 120W and 40V would be an RMS current of 3A in the combined output transistor set. If the outputs were to have hfe=3, they would demand 1 amp of base current which should be within the SOA of the MJE15032. Most likely the gain of the output transistors is going to be in the 10-50 which lowers the drive current needed. So my guess is that the MJE15032/3 should be fine if you can not get genuine NOS MJ15020/21 or MJ15018/9. I searched for other possibilities but found nothing else that could serve.

It going to be close for sure. I wonder if there are any TO-66 possibilities. I can pivot them a little and carefully drill a couple of new holes to accommodate them. They'll have a little better heat sinking than TO-220.

 

[Skywatcher]

I have never worked on one of these, so I compare it with the sansui AU-D907 (same as AU-919) which also has two complementary power transistors driven by a 1.5A driver (2SD382/2SB537) but with higher voltage rails (+/-56V). I see the schematics now and I understand your hessitation. This McIntosh uses +/-42 Volt rails, so to get 120W (comparable to the AU-919) it will have to pass about 35% more current than the AU-919. As a guideline, scaling up the 1.5A drivers in the AU-919 by 35% yields a little over 2A. The SOA for the MJ15020 shows almost 4A capability at 40V, while the MJE15032 is down to 2A at that voltage, unfortunately it is barely there.
EDIT:
The estimate avobe is a little sketchy.

For 120W and 40V would be an RMS current of 3A in the combined output transistor set. If the outputs were to have hfe=3, they would demand 1 amp of base current which should be within the SOA of the MJE15032. Most likely the gain of the output transistors is going to be in the 10-50 which lowers the drive current needed. So my guess is that the MJE15032/3 should be fine if you can not get genuine NOS MJ15020/21 or MJ15018/9. I searched for other possibilities but found nothing else that could serve.

Any thoughts on these in TO-3P? Yes obsolete but still available...

https://www.semicon.sanken-ele.co.jp/sk_content/2sa1693_ds_en.pdf
https://www.semicon.sanken-ele.co.jp/sk_content/2sc4466_ds_en.pdf

https://media.digikey.com/pdf/Data Sheets/ST Microelectronics PDFS/2STA1694.pdf
https://media.digikey.com/pdf/Data Sheets/ST Microelectronics PDFS/2STC4467.pdf

For reference...

https://pdf1.alldatasheet.com/datasheet-pdf/view/4800/MOTOROLA/MJ15018.html

 

[JoseHH]

Nice find! Big improvement relative to the TO-220 MJE pair we were discussing. In my opinion, because of the SOA figure, and the highest possible Vceo, these are the best of the two pairs:

https://media.digikey.com/pdf/Data Sheets/ST Microelectronics PDFS/2STA1694.pdf
https://media.digikey.com/pdf/Data Sheets/ST Microelectronics PDFS/2STC4467.pdf

Transition frequency is quite close, they are a complementary pair, and at 40V they are very close to the MJ15020/21. Also important, they will fit in the TO-3 location with almost no changes for mounting.

 

[Skywatcher]

Nice find! Big improvement relative to the TO-220 MJE pair we were discussing. In my opinion, because of the SOA figure, and the highest possible Vceo, these are the best of the two pairs:


Transition frequency is quite close, they are a complementary pair, and at 40V they are very close to the MJ15020/21. Also important, they will fit in the TO-3 location with almost no changes for mounting.

I ordered a few while I was there. They're pretty inexpensive too which surprised me. The TO-3P was developed to replace the TO-3 as a drop-in. The lead spacing is the same and you bend the leads right at the taper to fit the sockets. There's a diagram floating around here somewhere. The TO-220 also drops in a TO-66 socket the same way. If you look closely at the sockets, they have a "split" so you get 4 points of contact with the flat leads with plenty of "grip". I tin them anyways to help "round" them out. The metal tab is still the Collector so the center lead isn't needed for function. But you still need the 2nd screw to hold the socket in so I will ring term it and use it. The TO-3 mica will still work too making it convenient for the center lead.

Crossing fingers they work without any issues or oscillations.

Found the retrofit diagrams...

 

[JoseHH]

The parts you ordered are probably the closest match to what was in that amplifier. I doubt you need 230V parts for 42 volts rails but current handling is important. Yes that diagram is very useful. I am curious to know how well the parts you are getting perform. Please let us know.

 

[Skywatcher]

The parts you ordered are probably the closest match to what was in that amplifier. I doubt you need 230V parts for 42 volts rails but current handling is important. Yes that diagram is very useful. I am curious to know how well the parts you are getting perform. Please let us know.

The MJ15018/19 were also used as replacement drivers in these so I think we're "safe" enough in that respect. Even if they don't perform "optimally" (sonic), and have no functional "side-effects", it beats silence any day of the week. But I think they'll do fine. And since it is only one channel, they can be directly compared to the "original".

 

[Skywatcher]

Still waiting on the Digi-Key order. They are behind on getting orders processed, but was finally shipped yesterday. Plus the Holiday season has begun..

 

[Skywatcher]

OK a bit of an update. After a "mixup" at Digi-Key, the transistors finally arrived. There are a couple of caveats with these TO-3P to deal with. They are a plastic body with metal tab so using the one screw does not have a collector connection. I put a ring term on the collector lead to solve that. McIntosh uses fatter screws which means the holes in the transistors are just a bit too small. A little work with a Dremel and diamond burr to ream the holes did the trick. Only a little is needed so it went fast. Also the thickness of the transistors makes the one screw just long enough to catch enough threads. But it has enough to do the job. The metal tabs also need one corner ground off because at the angle it was too close for comfort to the heat sink fin. Again, only a few seconds at the grinder and it's fine. I chose to round off both corners which also closer matches the TO-3 insulators. When I get a chance I'll get the photos off my phone, onto my computer, and upload them.
Now onto the amp card. I found one transistor open on all 3 leads, but everything else tests fine so far. The part number is 132-143 (Q221) which crossed to an MPSD05 and is on the -43v power. Puzzling since the MPSD05 is only a 25v part I cannot find any anyways. However I do have BC337-40 on hand which looks very close in spec and high enough voltage. So the question is, will it work? Or are there any other suggestions for a substitute. Of course I have the compliment for the +43v side too.

 

[IvoCRO]

The 'other' side of Q221 should newer be on 0 V so 25 V is more than enough. However, if Q221 was damaged, I suspect that 'other' side of this transistor went towards ground. Check all transistors and diodes on driver board on right side of schematic. When I got MC2200 (which uses same driver boards as MC2120) R273 was blown, as well as few 'small' transistors, so check resistors too. Most of them can be checked without lifting them from board.

 

[Skywatcher]

The 'other' side of Q221 should newer be on 0 V so 25 V is more than enough. However, if Q221 was damaged, I suspect that 'other' side of this transistor went towards ground. Check all transistors and diodes on driver board on right side of schematic. When I got MC2200 (which uses same driver boards as MC2120) R273 was blown, as well as few 'small' transistors, so check resistors too. Most of them can be checked without lifting them from board.

All resistors check out as well as all other transistors on the Peak DCA75 tester. Later today I will move on to the diodes. But I will recheck R273 just to be thorough. Thanks.

 

[Skywatcher]

Here are a few pics of the TO-3P drivers.

Prepped and holes enlarged to fit the screws.


Mounted



The TO-3P leads on the underside of the sockets on the right. Snipped to the same length of the TO-3 leads. I gave them a twist so they miss the split to get all 4 points of contact. They went in a bit snug.

 

[Skywatcher]

Found shorted Stabister diodes. 070-046 (MZ2361 - Obsolete) in series with an 070-047 (1N4148 - got plenty). I don't see why two 1N4148 in series won't work. Sansui VD1212 (and a couple others) get replaced with two 1N4148 in series.
Any thoughts pro or con?

 

[jlovda]

I'm confused. The metal tab on the back of the plastic body of the transistor is NOT connected to the center pin? Why couldn't you just clip the center pin off and put the screw back in the hole without any contact with the transistor? The screw on the tab is not tight or strong enough to hold the part?

 

[w3jn]

SOmetimes the screw threads have separate wires attached to each one, and depend upon the transistor case to conduct across them. One could just wire across the bottom, but this is also a solution.

You can see this in the third photo of post #17. There's a resistor attached to one of the screw receptacles and a white wire attached to the other, on the upper right transistor.