I've got an old Adcom GFP-750 preamp that I bought used many years ago. I just got it out of storage and realized how much I like it. One thing that's always bothered me though is that a mod was made to the power supply before I obtained it, most likely fix a blown voltage regulator transistor.
I've included a couple pictures. One is a closeup of the power supply area modded, and the second is a page in the service manual addressing the issue modded.
It appears as though my unit has the IRF610 FET shown in Figure 1. The strange thing is the rest of the circuit on my board is that shown in Figure 2. It appears as though the FET should be replaced with the 2SC4793 BJT.
Does this configuration of my unit make any sense? If I do change the part, I'd like to mount it properly, as the original ones were (as the Motorola 7805 is mounted).
I've found 2SC4793 here:
Search
I don't, however know where to buy the proper mounting hardware.
Lastly, it seems to operate just fine. Should I just avoid fixing what isn't broken?
Any help would be greatly appreciated!
Thanks, Jim
I've included a couple pictures. One is a closeup of the power supply area modded, and the second is a page in the service manual addressing the issue modded.
It appears as though my unit has the IRF610 FET shown in Figure 1. The strange thing is the rest of the circuit on my board is that shown in Figure 2. It appears as though the FET should be replaced with the 2SC4793 BJT.
Does this configuration of my unit make any sense? If I do change the part, I'd like to mount it properly, as the original ones were (as the Motorola 7805 is mounted).
I've found 2SC4793 here:
Search
I don't, however know where to buy the proper mounting hardware.
Lastly, it seems to operate just fine. Should I just avoid fixing what isn't broken?
Any help would be greatly appreciated!
Thanks, Jim
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
The Toshiba 2SC4793 is a TO-220FP, the body is completely insulated, no hardware needed except a screw and maybe some heatsink goop. The circuit should be one or the other NOT a combo of the two.
Craig
Craig
Thanks for the reply!
That's what I was thinking. It makes no sense what was done. Somehow it works.
So you recommend removing the insulation pad, and simply using goop instead? I'm going to order the 2SC4793 BJT and goop.
That's what I was thinking. It makes no sense what was done. Somehow it works.
So you recommend removing the insulation pad, and simply using goop instead? I'm going to order the 2SC4793 BJT and goop.
Not sure. I'll open it up again when I get some time and check it. The heatsink it's attached to is probably 100 deg F or so, lightly warm to the touch.
I've serviced quite a few of these. When they have the mosfet power supply, and it's still working, I just leave it be, and I haven't had any fail in the field. I did have a few units where the body of the to-220 casing was not flat against the heatsink, and it was overheating. So I suspect a lot of the failures that Adcom was seeing—that prompted them to issue the revision to the power supply—was actually overheating of the mosfet, and not anything wrong with the design.
If you decide to change to the bipolar transistor. Study the schematic in the service bulletin closely, there are some subtle differences.
But if it's working already, I would keep the mosfet. But I suggest you just replace that flimsy nylon mounting screw with a metal bolt and insulating shoulder washer, and install a fresh thermal pad.
BTW, most of the heat dissipated to the heatsink actually comes from the LM7805 5V regulator for the logic circuits.
If you decide to change to the bipolar transistor. Study the schematic in the service bulletin closely, there are some subtle differences.
But if it's working already, I would keep the mosfet. But I suggest you just replace that flimsy nylon mounting screw with a metal bolt and insulating shoulder washer, and install a fresh thermal pad.
BTW, most of the heat dissipated to the heatsink actually comes from the LM7805 5V regulator for the logic circuits.
Thanks for the great information. Yea that nylon screw is a hack.
The problem with my unit is I don't have the FET circuit.
I have a bizarre combination of things. My board has R30/R31 on it, so it's a board made for the BJT circuit. The strange thing is all the mods for the BJT circuit are done manually. The mods are clearly not original work, and not very good either. You can see in the picture the diode added in parallel to R31.
Then add to this the FET part within the BJT circuit. 😱
The problem with my unit is I don't have the FET circuit.
I have a bizarre combination of things. My board has R30/R31 on it, so it's a board made for the BJT circuit. The strange thing is all the mods for the BJT circuit are done manually. The mods are clearly not original work, and not very good either. You can see in the picture the diode added in parallel to R31.
Then add to this the FET part within the BJT circuit. 😱
Last edited:
I felt the IRF610 and it does get hot to the touch after about 5 seconds or so. It's actually much hotter than the LM7805 in my circuit concoction.
Last question, anyone know where I can get this TO-220 package spacer? I need one to use the original screw through hole in the heatsink. Thanks!
I ordered the Toshiba BJT and mounting hardware.
I ordered the Toshiba BJT and mounting hardware.
I've never seen one of those type of attachment devices as displayed in your picture before. However it looks like the heatsink was drilled and tapped for a standard mounting option based on the nylon screw that is currently there.
If you're staying with a TO220 package (IRF610) that needs to be insulated then use one of the plastic washers and a steel bolt. The Mouser part 534-3049 insulator will work for 220 packages and allows up to a #6 bolt IIRC.
However if you're installing a C4793 as outlined in the service manual then someone already pointed out it'll be a 220FP and will not require any insulation from the heatsink. Just put some thermal paste on it and bolt it down using the same hole that is attaching the mosfet now.
If you're staying with a TO220 package (IRF610) that needs to be insulated then use one of the plastic washers and a steel bolt. The Mouser part 534-3049 insulator will work for 220 packages and allows up to a #6 bolt IIRC.
However if you're installing a C4793 as outlined in the service manual then someone already pointed out it'll be a 220FP and will not require any insulation from the heatsink. Just put some thermal paste on it and bolt it down using the same hole that is attaching the mosfet now.
Thanks Chamberman. Yea I think the extender is a no-go. Doesn't really matter. I ordered a set of M3 hex screws to make it look pretty. I'm going to change to the C4793 BJT. I was planning on just using a silicone insulator pad, but if thermal paste works better, maybe that's a better idea. How would this work?
https://www.amazon.com/100Gram-Ther...ter+Thermal+Paste+white&qid=1608917420&sr=8-3
Sorry I'm a newbie at transistor mounting...
https://www.amazon.com/100Gram-Ther...ter+Thermal+Paste+white&qid=1608917420&sr=8-3
Sorry I'm a newbie at transistor mounting...
If you're going with the TO-220FP transistor then I'd not use a silicone pad. The thermal conductance will be worse with the pad than it will with nothing but a thin smear of thermal compound behind it. The compound you linked to should work fine.