SX-1980 power supply problems

[txturbo]

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.

Good catch.. I'll add them in.

I was thinking about the 80V transistor connector's polarization issue. Maybe it's best to make the connection on the board and solder wires onto the transistor. I know it's a pain to replace but at least the connection to the board can be polarized. I am not sure what would happen if the transistor was connected backwards. Thoughts?

 

[EchoWars]

A drop of nail polish on the connector to indicate the 'up' position would do. It isn't hard to see whether the connector is on backwards or not...follow the wires from the transistor to the board. Once you 'do' one, you'll see what I mean.

 

[rcs16]

12. The board has an internal ground plane. This is a bit subtle when you look at the layout.

Okay i am not sure what you mean, so the pcb is how many layers of copper? Internal usually means more than two layers. Maybe you are referring to copper pours for the ground net?
Make sure that the ground returns on pins 10, 21 are separated and they meet at the common point ground which is situated at the big ecap bus bar and where the transformer "0" wires meet.
I see that they are!!

EW are you going over the schematic comparing against the original?

 

[txturbo]

Okay i am not sure what you mean, so the pcb is how many layers of copper? Internal usually means more than two layers. Maybe you are referring to copper pours for the ground net?
Make sure that the ground returns on pins 10, 21 are separated and they meet at the common point ground which is situated at the big ecap bus bar and where the transformer "0" wires meet.
I see that they are!!

EW are you going over the schematic comparing against the original?

The board is actually 4 layers. The top and bottom are for signals and there are two internal layers that are typically power and ground. Because this is a power supply board the power layer has been eliminated. The ground layer connects everything in the schematic that is ground together.

The software checks the schematic against the layout to see that everything is connected properly.

The picture of the layout a few posts back omits the ground plane. When it's activated in the view on the layout you can't see any of the traces.

I'll post a picture of the latest layout in a bit. I need to get to work and slurp some coffee!

 

[txturbo]

Latest layout

Here is the latest layout, it should have everything in it. The dropping resistor pads are on the board for the relay, if they are needed the trace can be cut and the resistor installed.

The green layer is the bottom copper.
The red layer is the top copper.
Yellow is the silkscreen on the top.
The ground plane is omitted from the view.

There is a legacy trace near R222 that will be removed, it connects to nothing. (Fixed)

Also something I should have mentioned before... C1 and C2 were physically soldered to the posts 20, 21 and 23. They have been moved to the board.

 

[satellite65]

Enjoying this thread...

Just curious, are you using ExpressPCB?
I design test fixtures at work and have used them several times.
Their boards are top quality.

 

[txturbo]

Enjoying this thread...

Just curious, are you using ExpressPCB?
I design test fixtures at work and have used them several times.
Their boards are top quality.

We have used them along with a couple others and a local place here in Austin.

 

[hrballenger]

Great work One of my biggest frustrations working with the original AWR-154 boards is the lack of a good top-side silkscreen to reference. Any possibility of working one up for the original boards while you have all the resources right there ? Would be a HUGE help for us oldsters with failing eyesight and shorter fuses on our frustration levels.

 

[txturbo]

Great work One of my biggest frustrations working with the original AWR-154 boards is the lack of a good top-side silkscreen to reference. Any possibility of working one up for the original boards while you have all the resources right there ? Would be a HUGE help for us oldsters with failing eyesight and shorter fuses on our frustration levels.

Are you looking for a print out that you can use as a reference?

 

[txturbo]

Time to make some boards...

Unless anyone has any further comments or suggestions I am going to get a board built and give it a go... Thanks for all of the help!

As soon as I have one in hand I'll post it up

 

[EchoWars]

Need to post a list of the oddball parts used...the LY2 relays, the thermal fuse, the safety cap on the soft-start relay, part #'s for the connectors you chose, and (since I did not see what you planned on using) the soft-start resistor.

Wanted to ask what lead spacing was used for the radial caps on C204 and C205. I use a Panasonic EEU-ED2C221S 220µf 160V (Digikey #P13512-ND). 18mm diameter x 25mm height, with a 7.5mm lead spacing. I might note here that it may be a good idea to have a double-pad for these two radial caps to allow a 7.5mm lead spacing and another pad for 10mm lead spacing.

Matter of fact, the best answer here is to set this up for a couple of radial snap-in caps for the 10mm spacing. For a 220µf 160V cap, a 22mm diameter is very common. I don't know if the caps would have to be offset to fit two 22mm caps in there or not...software should tell you if it will do or not.

Example snap-in parts:
Mouser: #647-LGJ2C221MELZ Nichicon LGJ 105°C 220µf 160V 22mm x 20mm
#598-SLP331M160A1P3 Cornell Dubilier SLP 105°C 330µf 160V 22mm x 25mm
#598-SLPX471M160A3P3 Cornell Dubilier SLP 105°C 470µf 160V 22mm x 30mm
#647-LGN2C331MELY25 Nichicon LGN 105°C 330µf 160V 22mm x 25mm
...and more.

Digikey: #LGJ2C221MELZ-ND Nichicon LGJ 220µf 160V 105°C 22mm x 20mm
#P13944-ND Panasonic T-UP 470µf 160V 105°C 22mm x 27mm
#493-2621-ND 270µf 160V Nichicon LGU 105°C 22mm x 25mm
#338-1489-ND 330µf 160V Cornell Dubilier SLP 105°C 22mm x 25mm
and more...

Newark: #97M5607 Cornell Dubilier SLP 470µf 105°C 160V 22mm x 30mm

Bottom line is, any two-terminal snap-in cap has a lead spacing of 10mm. And if a 22mm cap can be made to fit, then any snap-in with a capacitance rating from 220µf to 470µf and a voltage rating of at least 160V, and a maximum height of about 30mm will work.

Hole diameter is much larger for a snap-in cap than it is for a leaded radial cap (1.5mm leads on snap-in caps).

Edit: Ya know, I typed all this BS and then looked at the foil layout again. Looks like you already have large holes for the 'radials replacing the three axial' caps. If that's the case, fill us in on the maximum diameter that can be used, and if the hole spacing is the proper 10mm. Whatever the answer, the maximum height will still have to be around 30mm. And if this is set up for snap-in caps as I suspect and can handle 22mm diameter, I'd not bother with holes for a 7.5mm leaded radial cap.

Shoulda looked closer first.:stupid:

 

[txturbo]

Update on the odd parts

All the caps pads are selected for either the original style axial or a 10MM snap in for flexibility. The part we planed to use was 25MM high for the 330uF and 20MM for the 220uf.

The relays are both LY2-0-DC24

The thermal fuse is a G4A01110C It's rated 1 deg higher than the original.

The safety cap is a DEJF3E2103ZB3B

The connectors are Molex Euro Plugs. Rated for 18A per contact on the female side and 15A on the male, screw down from the top with the wire entering at 90 deg and fully polarized. Thees are actually really nice.

The soft start resistor is a WPRT20AB-3R3JB270

Need to post a list of the oddball parts used...the LY2 relays, the thermal fuse, the safety cap on the soft-start relay, part #'s for the connectors you chose, and (since I did not see what you planned on using) the soft-start resistor.

Wanted to ask what lead spacing was used for the radial caps on C204 and C205. I use a Panasonic EEU-ED2C221S 220µf 160V (Digikey #P13512-ND). 18mm diameter x 25mm height, with a 7.5mm lead spacing. I might note here that it may be a good idea to have a double-pad for these two radial caps to allow a 7.5mm lead spacing and another pad for 10mm lead spacing.

Matter of fact, the best answer here is to set this up for a couple of radial snap-in caps for the 10mm spacing. For a 220µf 160V cap, a 22mm diameter is very common. I don't know if the caps would have to be offset to fit two 22mm caps in there or not...software should tell you if it will do or not.

Example snap-in parts:
Mouser: #647-LGJ2C221MELZ Nichicon LGJ 105°C 220µf 160V 22mm x 20mm
#598-SLP331M160A1P3 Cornell Dubilier SLP 105°C 330µf 160V 22mm x 25mm
#598-SLPX471M160A3P3 Cornell Dubilier SLP 105°C 470µf 160V 22mm x 30mm
#647-LGN2C331MELY25 Nichicon LGN 105°C 330µf 160V 22mm x 25mm
...and more.

Digikey: #LGJ2C221MELZ-ND Nichicon LGJ 220µf 160V 105°C 22mm x 20mm
#P13944-ND Panasonic T-UP 470µf 160V 105°C 22mm x 27mm
#493-2621-ND 270µf 160V Nichicon LGU 105°C 22mm x 25mm
#338-1489-ND 330µf 160V Cornell Dubilier SLP 105°C 22mm x 25mm
and more...

Newark: #97M5607 Cornell Dubilier SLP 470µf 105°C 160V 22mm x 30mm

Bottom line is, any two-terminal snap-in cap has a lead spacing of 10mm. And if a 22mm cap can be made to fit, then any snap-in with a capacitance rating from 220µf to 470µf and a voltage rating of at least 160V, and a maximum height of about 30mm will work.

Hole diameter is much larger for a snap-in cap than it is for a leaded radial cap (1.5mm leads on snap-in caps).

Edit: Ya know, I typed all this BS and then looked at the foil layout again. Looks like you already have large holes for the 'radials replacing the three axial' caps. If that's the case, fill us in on the maximum diameter that can be used, and if the hole spacing is the proper 10mm. Whatever the answer, the maximum height will still have to be around 30mm. And if this is set up for snap-in caps as I suspect and can handle 22mm diameter, I'd not bother with holes for a 7.5mm leaded radial cap.

Shoulda looked closer first.:stupid:

 

[jheu02]

I've been reading this thread with interest. My one thought is...you've got some great brains helping guide this new board. Though I know the desire to get your personal unit working again, going to these lengths shouldn't be rushed. You'll save yourself potentially many later issues by just taking a quick breather to let other's mull it over.

Kind of like writing a term paper...the writer is so focused on the paper he doesn't see the simple grammatical or spelling errors that others looking at the work catch.

 

[hrballenger]

Yes, something a bit better than the one in the service manual or the board itself.

Are you looking for a print out that you can use as a reference?

 

[txturbo]

Yes, something a bit better than the one in the service manual or the board itself.

The problem is that the new board layout is not exactly the same as the old one.

Specifically the pads have been changed to allow the new transistors to be properly oriented. Also the silkscreen won't match exactly because things have been moved around a bit.

 

[jheu02]

I think he was asking if, as you started this process, your first rendition was of the ORIGINAL layout, before you started making changes and moving things.

 

[EchoWars]

I also see that since Q1, Q2, Q3, and Q4 are padded on the PC board with the leads in a tight row, it's going to be important to state whether this is set up for a 'center base' part or a 'center collector' part. Mentioned because 95% of Japanese transistors are center-collector parts. When I did the first current source, I just happened to use the center-base ZTX transistors. So you can see there might be some confusion unless the intended parts are clearly stated.

Would have been nice to lay out as many of the TO-92 transistor pads in a 'circular' arrangement as possible, much like Pioneer does on most of their gear. This allows easy use of either center-base or center collector parts. But, that would require a major revision to the PC board, so I'll simply say that stating the intended components will be necessary. (would be a big help to silkscreen on the PC board the lead ID's [E - B - C].

 

[txturbo]

I think he was asking if, as you started this process, your first rendition was of the ORIGINAL layout, before you started making changes and moving things.

We never did capture the original layout exactly...

Just to give you an idea of the process we are using:

1. Capture the schematic in the program, including any changes.

2. Create an initial BOM including sourcing and substituting of components.

3. Measure the board.

4. Create any parts that need to be added to the library.

5. Pull parts form the library and place them on the board as closely as possible to the original locations.

6. Measure traces and duplicate as close as possible to the original.

7. Check each trace against the schematic (this is done in the program and manually).

8. Add any special features, heat sinks, ground pads, mounting holes....

9. Initial rev complete.

9. Do a complete review..

10. Modify

11. Do a review.

12. Modify

13. Order initial run and hand stuff board.

14. Test

15. Modify if necessary on board.

16. Test

17. Order final run.

18. If low volume build by hand, if high volume send to board house.


We have done this over 250 times on boards we use for equipment that we refurbish. Most of the time we get it right on the first build but sometimes we f it up.

I also see that since Q1, Q2, Q3, and Q4 are padded on the PC board with the leads in a tight row, it's going to be important to state whether this is set up for a 'center base' part or a 'center collector' part. Mentioned because 95% of Japanese transistors are center-collector parts. When I did the first current source, I just happened to use the center-base ZTX transistors. So you can see there might be some confusion unless the intended parts are clearly stated.

Would have been nice to lay out as many of the TO-92 transistor pads in a 'circular' arrangement as possible, much like Pioneer does on most of their gear. This allows easy use of either center-base or center collector parts. But, that would require a major revision to the PC board, so I'll simply say that stating the intended components will be necessary. (would be a big help to silkscreen on the PC board the lead ID's [E - B - C].

How about EBC with some test point copper pads spread out a bit to avoid shorting.

 

[Overundr1]

Would have been nice to lay out as many of the TO-92 transistor pads in a 'circular' arrangement as possible, much like Pioneer does on most of their gear. This allows easy use of either center-base or center collector parts. But, that would require a major revision to the PC board, so I'll simply say that stating the intended components will be necessary. (would be a big help to silkscreen on the PC board the lead ID's [E - B - C].


What EW suggested
-Lee

 

[txturbo]

Would have been nice to lay out as many of the TO-92 transistor pads in a 'circular' arrangement as possible, much like Pioneer does on most of their gear. This allows easy use of either center-base or center collector parts. But, that would require a major revision to the PC board, so I'll simply say that stating the intended components will be necessary. (would be a big help to silkscreen on the PC board the lead ID's [E - B - C].


What EW suggested
-Lee

Done