SX-1980 power supply problems
- Thread starter txturbo
- Start date
Of course, the two contacts on the soft-start get run in parallel. Just to make sure we're on the same page. 
And a spark-killer across the soft-start relay contacts would be nice as well...a .01uf safety cap in series with a 100 ohm 1/2W resistor.
Caps:
440V
BC2698-ND
399-9538-1-ND
250V
490-9486-ND
490-9487-ND
490-9564-1-ND
Digikey #'s all.
And a spark-killer across the soft-start relay contacts would be nice as well...a .01uf safety cap in series with a 100 ohm 1/2W resistor.
Caps:
440V
BC2698-ND
399-9538-1-ND
250V
490-9486-ND
490-9487-ND
490-9564-1-ND
Digikey #'s all.
Last edited:
I don't think you should rush this to the board house yet, let's let the idea sit and percolate a bit - I myself have many issues tugging at me, even though I consider this an important issue to address.
I need to sit down, with my spare awr-154 board and THINK of all the hassles they have hit me with over the years. We've covered quite a few, and for example you surprised me with the 3.3 ohm 20w resistor (which I haven't followed up on yet either) being available.
Like, what about all the other to-220 transistors on the board that are series pass regulators, what is the intent for cooling them?
It's a heck of a lot easier to correct computer files and paper designs, than deal with the tears that come with spending $$ on immutable hard assets.
I've been through this before, and I DO acknowledge the saying "shoot the engineers and just SHIP the damn thing!!"...
Engineering innovation is fun, but it is more about getting all the pertinent and applicable details under control.
I need to sit down, with my spare awr-154 board and THINK of all the hassles they have hit me with over the years. We've covered quite a few, and for example you surprised me with the 3.3 ohm 20w resistor (which I haven't followed up on yet either) being available.
Like, what about all the other to-220 transistors on the board that are series pass regulators, what is the intent for cooling them?
It's a heck of a lot easier to correct computer files and paper designs, than deal with the tears that come with spending $$ on immutable hard assets.
I've been through this before, and I DO acknowledge the saying "shoot the engineers and just SHIP the damn thing!!"...
Engineering innovation is fun, but it is more about getting all the pertinent and applicable details under control.
Well, the +/-34V pass trannies run warm, not hot, with the dinky little heatsinks they already have. And the plan is, with this board, to move the +/-80V transistors to the large output heatsinks. This leaves the 8V and the 13V heatsinks.
I believe the plan is to use these heatsinks for the 8V and 13V pass transistors, though we might want to check with turbo on that. And although it isn't mentioned, I'd strongly suggest dropping a bit more voltage to the 13V regulator by changing R202 from a 3.3 ohm 1/2W resistor to a 22 ohm 3W resistor, and mounting the resistor a respectful distance away from Q203 so that the additional heat isn't added to the transistors' battle.
Post #27 in this thread has a pic of a completely modded 'stock' board. Current sources, 80V trannies moved to the big sinks, new heatsinks for the 8V and 13V trannies, radial caps for C204 and C205, and the 22 ohm 3W resistor for R202. The end result is much more 'open' and airflow is improved mightily over the bone-stock setup.
The fly in the ointment, as it were, is Molex discontinuing the nifty TO-220 sockets I was using. I liked having a very quick way of disconnecting these transistors without unbolting them, should the PC board need attention. I've recently found the Molex part #22-02-7033 (Mouser #538-22-02-7033), but I've yet to sample one to check its suitability.
I believe the plan is to use these heatsinks for the 8V and 13V pass transistors, though we might want to check with turbo on that. And although it isn't mentioned, I'd strongly suggest dropping a bit more voltage to the 13V regulator by changing R202 from a 3.3 ohm 1/2W resistor to a 22 ohm 3W resistor, and mounting the resistor a respectful distance away from Q203 so that the additional heat isn't added to the transistors' battle.
Post #27 in this thread has a pic of a completely modded 'stock' board. Current sources, 80V trannies moved to the big sinks, new heatsinks for the 8V and 13V trannies, radial caps for C204 and C205, and the 22 ohm 3W resistor for R202. The end result is much more 'open' and airflow is improved mightily over the bone-stock setup.
The fly in the ointment, as it were, is Molex discontinuing the nifty TO-220 sockets I was using. I liked having a very quick way of disconnecting these transistors without unbolting them, should the PC board need attention. I've recently found the Molex part #22-02-7033 (Mouser #538-22-02-7033), but I've yet to sample one to check its suitability.
I have most of this design captured in ltspice, if it is of any use. It will calculate everything that one would need to know electrically. It is not thermal modeling however.
I like the fact that this is the first Pioneer pcb re-design that I am aware of.
If it was my design, it would include short circuit current limiting, not by means of a fuse or a fusible link = smoke. Even OVP on the tuner supplies is doable too. I think we have heard of some PAxxxx chips being fried because of a OV fault on the PS. Replace secondary regulators with newer IC's which also have limiting characteristics.
But that a choice of the designer/team and not mine to make.
I can offer to review the gerber/drill fab data and I guess I could do some sim if it is necessary. But it is some simple algebra to get the numbers you need to work with. Not sure if you want to publicly release this fab data, so we can hook up offline if you wish for my review.
Not too sure about the KK conn 22-02-7033 use for TO-220 leads which are more flatter than a 25mil sq post, got to give it a try and test for contact R. I guess you could solder 25mil sq posts to the TO-220 leads as a hack.
I like the fact that this is the first Pioneer pcb re-design that I am aware of.
If it was my design, it would include short circuit current limiting, not by means of a fuse or a fusible link = smoke. Even OVP on the tuner supplies is doable too. I think we have heard of some PAxxxx chips being fried because of a OV fault on the PS. Replace secondary regulators with newer IC's which also have limiting characteristics.
But that a choice of the designer/team and not mine to make.
I can offer to review the gerber/drill fab data and I guess I could do some sim if it is necessary. But it is some simple algebra to get the numbers you need to work with. Not sure if you want to publicly release this fab data, so we can hook up offline if you wish for my review.
Not too sure about the KK conn 22-02-7033 use for TO-220 leads which are more flatter than a 25mil sq post, got to give it a try and test for contact R. I guess you could solder 25mil sq posts to the TO-220 leads as a hack.
Last edited:
Just to be clear...
Any changes at this point will be simple to do and I am in no rush other than I would like to see my 1980 working again.
We had planned on soldering on pigtails with connectors that plug into the board for the +/-80V parts. We looked at the connectors for the part and there does not seem to be a great set of choices. With connectors on the end of the pigtail and board we have many to choose from.
Any changes at this point will be simple to do and I am in no rush other than I would like to see my 1980 working again.
We had planned on soldering on pigtails with connectors that plug into the board for the +/-80V parts. We looked at the connectors for the part and there does not seem to be a great set of choices. With connectors on the end of the pigtail and board we have many to choose from.
I don't have the parts list at home so I can't check on the heat sinks that we planned on using for the 8v and 13v. I know they are a bit larger than the original ones.
Moving R202 and changing it to 22 ohm 3W is no problem.
The board is already set up for axial or radial caps.
We put the connector on the board for the 80V's rather than on the part. Do you see any downside to this?
If I had to chose the next part to address it would be the protection IC. It would be great to make it out of discrete parts that can be obtained. I did search for the original IC and it looks like it's available in decent qty's so I figured it's not worth messing with. If necessary it could be moved from the old board.
I'd say leave PA3004 as is too and have a spare. I have heard of one failure on this site.
I am going to re-use a HA12002 on the Sansui G-7500 mod using a sx-1050 PS, just to make it easy for me since I do not have a 1050 protect pcb.
I just noticed that the 5.0V supply, shown as wire "l" on the big schem. APC AWX-124,pin2, it relies on the V drop of R209(39ohm,1/2W) to drop the reg ~8V down, relying on an avg of 80mA of load I. This is an example of a BAD design, could have used a 5.1V zener on it !! Accumlated tolerances all over the place.
Makes little sense, this design, since it powers a xtal reference for the APC. It is a sensitve ckt, it uses feed-thru caps in a shielded cage!!
I am going to re-use a HA12002 on the Sansui G-7500 mod using a sx-1050 PS, just to make it easy for me since I do not have a 1050 protect pcb.
I just noticed that the 5.0V supply, shown as wire "l" on the big schem. APC AWX-124,pin2, it relies on the V drop of R209(39ohm,1/2W) to drop the reg ~8V down, relying on an avg of 80mA of load I. This is an example of a BAD design, could have used a 5.1V zener on it !! Accumlated tolerances all over the place.
Makes little sense, this design, since it powers a xtal reference for the APC. It is a sensitve ckt, it uses feed-thru caps in a shielded cage!!
What connector?
No way I'd be fooling with soldering posts into the transistor. Way too hackish.
But, for grins, I checked a 3-pin computer fan, and the connector has 2.54mm spacing and the one I had here grips the transistor lead nicely. So...order a 12" computer fan extension (or some such) and clip off any unneeded wire (the wires length needed to move the 80V transistors is 11").
http://www.newegg.com/Product/Product.aspx?Item=N82E16812189006
http://www.newegg.com/Product/Product.aspx?Item=N82E16812119148
Last edited:
elvi
Super Member
No way I'd be fooling with soldering posts into the transistor. Way too hackish.
But, for grins, I checked a 3-pin computer fan, and the connector has 2.54mm spacing and the one I had here grips the transistor lead nicely. So...order a 12" computer fan extension (or some such) and clip off any unneeded wire (the wires length needed to move the 80V transistors is 11").
We have an application that draws a bit of current but was also in a hot area.
We at one time used a similar connector but found it lost its spring and gained resistance.
We switched to a Molex style.
I personally lobbied for hard soldered pigtail to the VReg and a Molex on the other.
I think hard soldered and shrink wrapped high strand count color coded wire with a Molex is a respectable and very serviceable solution.
Watching this with much anticipation.
Q210 and Q211 are actually connectors.
The plan was to have wires soldered directly to the transistors and then plug them into the board.
That part I got. I was asking what connector you had planned to use. If it is one where the wires are removed individually, I'd prefer to solder the wires to the board and use some sort of 3-pin connector at the transistor, such as the computer fan connector I mentioned in the last post.
Got it.. The connector we had planed on using on the board is similar to the fan connector. So I guess we could just come up with a cable that is female on both ends.
Do you think there will be enough clearance between the bottom cover and the connector?
In any case if need be the entire thing can just be hard wired.
Ahhh...I C. You'd have to 'train' the wires with a bit of a bend, but clearance for the connector is not an issue.
But...
The real concern is that, if the wires are soldered to the transistor, then replacing the transistor gets to be a real tedious PITA (solder, shrinkwrap, rinse, repeat) if the only removable connection is at the PC board. And having a connector at the transistor AND the PC board is adding yet another mechanical interface. Best to use one connector at the transistor and solder the wires to the PC board.
But...
The real concern is that, if the wires are soldered to the transistor, then replacing the transistor gets to be a real tedious PITA (solder, shrinkwrap, rinse, repeat) if the only removable connection is at the PC board. And having a connector at the transistor AND the PC board is adding yet another mechanical interface. Best to use one connector at the transistor and solder the wires to the PC board.
Seems simple enough.. For the wire connections at the board do you think the wires should be soldered to pins or directly soldered to the board?
Updated
OK... I think it's getting close.
We worked on the changes today and will take one last look at it tomorrow.
I did think about a small potential problem. The connector for the 80V transistors is not polarized, this could cause a big problem :tears: Suggestions?
Other than that I think we have just about everything covered. I know there is potential for more advanced changes to the circuit but that involves risk.
Basically everything here has already been tested and proven to work.
Items updated:
1. 22 ohm 3W resistor for R202, moved to open real estate.
2. Soft-start and speaker protection relays are now LY2-0-DC24, powered from C203 +.
3. Removed R102 and R308.
4. Spark killer added across the soft start contacts. .01uf and 100 ohm 1/2W.
5. Contacts for soft start in parallel.
6. All wires are now on connectors for easy install and removal with screw terminals. Screws are on the top of the connector with the wires entering the connector at the side. The entire connector plugs into the board with the wires attached.
7. 80V reg transistors have been moved to the large heat sink. Wires to be soldered to board with female connector at transistor.
8. The large electrolytic caps have mounting provisions for axial or radial.
9. The current source mod has been integrated into the board.
10. All of the substitute devices have been orientated with pad layouts that match the pinout of the device.
11. The board has been silkscreened with components id.
12. The board has an internal ground plane. This is a bit subtle when you look at the layout.
13. Q203 and Q205 have larger heat sinks. The heat sinks use pins that are soldered to a large isolated pads on the board.
14. The board can be built from 100% new components, most of them from Mouser.
Latest schematic: ( I forgot to take a snapshot of the layout )
OK... I think it's getting close.
We worked on the changes today and will take one last look at it tomorrow.
I did think about a small potential problem. The connector for the 80V transistors is not polarized, this could cause a big problem :tears: Suggestions?
Other than that I think we have just about everything covered. I know there is potential for more advanced changes to the circuit but that involves risk.
Basically everything here has already been tested and proven to work.
Items updated:
1. 22 ohm 3W resistor for R202, moved to open real estate.
2. Soft-start and speaker protection relays are now LY2-0-DC24, powered from C203 +.
3. Removed R102 and R308.
4. Spark killer added across the soft start contacts. .01uf and 100 ohm 1/2W.
5. Contacts for soft start in parallel.
6. All wires are now on connectors for easy install and removal with screw terminals. Screws are on the top of the connector with the wires entering the connector at the side. The entire connector plugs into the board with the wires attached.
7. 80V reg transistors have been moved to the large heat sink. Wires to be soldered to board with female connector at transistor.
8. The large electrolytic caps have mounting provisions for axial or radial.
9. The current source mod has been integrated into the board.
10. All of the substitute devices have been orientated with pad layouts that match the pinout of the device.
11. The board has been silkscreened with components id.
12. The board has an internal ground plane. This is a bit subtle when you look at the layout.
13. Q203 and Q205 have larger heat sinks. The heat sinks use pins that are soldered to a large isolated pads on the board.
14. The board can be built from 100% new components, most of them from Mouser.
Latest schematic: ( I forgot to take a snapshot of the layout )
Guess I should have been clearer about these two resistors...
Since the voltage at C203 is unregulated, the possibility exists that we might end up with a value that needs to be padded down a notch to keep the relays happy. I'd leave the positions for the two resistors in place. If the resulting voltage at C203 is reasonably close to 24V, then the resistor positions can be replaced with a zero-ohm jumper. If the voltage is much higher than 25V, a small resistor can be used to drop a volt or two.
Sorry...I was thinking this but did not articulate it. My bad. I'd set the pads up for a 1/2W resistor, as in a worst-case scenario I can't imagine needing anything larger.