Jump to content

Troubleshooting midiphy SEQ v4+


latigid on
 Share

Recommended Posts

Sort of answered my own question there.
If anyone has any advice on how to desolder the chip, what temperature to use, should I use flux or solder paste, it would be great.

I also have kapton tape for nearby components.

Link to comment
Share on other sites

  

6 hours ago, Smithy said:

So I've been getting erratic behavior with the LEDs on the switches of the JA PCB.
I wonder if i made the same mistake soldering the wrong type resistor network?
The markings on the IC at RN1 are 4816P LF1-103 C1809
instead of 4816P-T02-103LF.

I blame myself for not checking the BOM on midiphy and solely relying on the video!

Thankfully i have a hot air station, flux and solder paste so should be able to remedy this issue.

 

Yup, that's a common one! Actually Peter shows you the Mouser part number in the video but of course it is not 100% obvious and the part numbers are fairly similar.

If you used leaded solder, you should be able to remove it easily. Flux can help but it is not 100% necessary. Normally I try for about 300°C and a relatively high airflow for this. If you need higher temperature it will probably be okay for the RN, but you'll need to take care of the other chips, MEC switches and encoder.

When removing the chip, you might need to wiggle it to get the solder moving. When I do it, I always try to slide the chip laterally along the leads i.e. towards/along the long sides of the pads. This tends to keep the PCB pads intact and doesn't bend the chip leads, hopefully also the leads don't get too messy with solder blobs.

When done, if necessary clean up the PCB pads with flux/braid to make the surface flat again. If necessary do the same with the chip. If it's not possible to easily get things flat, you can add flux and use the reflow hot air station again.

Good luck!

Link to comment
Share on other sites

So I removed the chip fine but ended up lifting the pad for pin 1 when trying to solder the correct 10k resistor network. 

I had problems getting the solder to take to that pad when aligning the chip - I should have used the chisel tip instead of the bevel to apply solder to the corner pads first.I usually only use the bevel tip for drag soldering. 

I applied some flux also and that must have been enough to lift the pad. 

Is there anywhere I can solder a bridge between pin 1 and a pad it connects to?

 IMG_20210510_155627-1.thumb.jpg.fea1f4a93964c78049011ee7d09fcac9.jpg

Link to comment
Share on other sites

28 minutes ago, latigid on said:

No need to replace that pad, pin 1 does not connect to anything.

Best,
Andy

Can't believe I lucked out! :grin: 

I usually use a temp of 350 degrees C for soldering including drag soldering.

After this I would recommend using a temp closer to 300 degrees for SMT parts!

 

Thanks a lot for your help Andy.

The encoder, switches and LEDs are working beautifully.

Edited by Smithy
Link to comment
Share on other sites

Something has me stumped with the build guide video part @ 58 mins

 

Peter is soldering BAT54 diodes to the LeMEC pcb. 

I don't have any markings for BAT54 diodes on my LeMEC pcb. My PCB version is 1.0.

I only have markings for BAT54 diodes on LeMEC RH PCB of which I have version 1.3_R.

Here is the LeMEC pcb I have with the backside showing:

IMG_20210513_220214-1.thumb.jpg.41f1d23d0d1b8e5c21c0272dad98bb16.jpg

 

And heres the backside of the LeMEC RH PCBIMG_20210513_220159-1.thumb.jpg.dc3338bb100d5cd5ce5028fd900b3135.jpg

 

Is it okay to ignore the bat54 diodes on the LeMEC board and proceed?

Thanks.

Link to comment
Share on other sites

@Smithy as a note, we only sell the latest, matching revisions in the shop - the shown V1.0 is a "Neaderthal" :) one from the beta-test days, the first revision of the SEQ v4+ :). I hope Andy can tell you a bit more about it, otherwise we can just send you a new version.

Best regards,
Peter

Link to comment
Share on other sites

You will probably get bad ghosting without the BAT54s. Short of ordering a new board, just "stack" the BAT54s 1:1 on top of the BC818s, so with the text still facing up for both transistor and diode.

Please stuff R17 (any value up to 10k) or simply bridge it. C6 and the nearby diode can be left unpopulated.

Edited by latigid on
Link to comment
Share on other sites

Thanks for the prompt replies!

I must be insane (I guess we all know that right? ;) ) As I have went ahead and soldered 3 bat54 diodes on top of the BC818s.

 

It's not easy getting the solder to reach the pins of the diodes but I'm getting there!

5 more to go.

Link to comment
Share on other sites

On 5/14/2021 at 6:22 AM, latigid on said:

I just hope that it wasn't 3x BAT54 on top of each transistor ;-)

This approach works best I found when the solder has burned through its flux. If it's too hard, you could consider a fleck of copper tape or a strand of wire.


Haha not 3 on top of each other no!

I'm short 2 components from the BOM, they may have not been finalized when I orderered back in the stone age as Peter would say! ;)
I'm short the 4816P-1-103LF Resistor network for RN5 on the LE MEC RH PCB so I presume its needed?

Also i'm missing 929700-01-36-RK - the longer Pin headers. I believe these need to be about 8mm long from the top of the plastic to the top of the Pin itself (when standing upright).

I think I've found some equivalents on RS, could you double check Andy if you dont mind?

https://ie.rs-online.com/web/p/pcb-headers/6812310/

I'm going by dimension "B" in the datasheet:
https://docs.rs-online.com/36e1/0900766b8135fed6.pdf

Edited by Smithy
RN5 instead of RN2
Link to comment
Share on other sites

There are two types of SMT resistor networks in the SEQ v4+. The bussed one on the JA board and two isolated on lemec_R to serve as current limiting for the activity matrix. Just use the product page to make sure you're ordering 8x resistors in a 16-pin package. If there are 15 resistors, that's the bussed one. (Seems like your RS part is good.)

(There are also bussed THT arrays on lemec boards)

For the pin headers, even normal length ones could work but they tend not to make good enough contact. Adding up the thicknesses:

connector 5mm
PCB 1.6mm
inter-PCB spacing 3mm

- plastic thickness 2.8mm

6.8mm of pin
 

So I would suggest 826630, or actually 826935 as it matches the tin plating better. The 8mm one that you found would probably still work, just be aware that if it sticks through the header too much, it might contact the case.

 

Link to comment
Share on other sites

13 minutes ago, latigid on said:

There are two types of SMT resistor networks in the SEQ v4+. The bussed one on the JA board and two isolated on lemec_R to serve as current limiting for the activity matrix. Just use the product page to make sure you're ordering 8x resistors in a 16-pin package. If there are 15 resistors, that's the bussed one. (Seems like your RS part is good.)

(There are also bussed THT arrays on lemec boards)

For the pin headers, even normal length ones could work but they tend not to make good enough contact. Adding up the thicknesses:

connector 5mm
PCB 1.6mm
inter-PCB spacing 3mm

- plastic thickness 2.8mm

6.8mm of pin
 

So I would suggest 826630, or actually 826935 as it matches the tin plating better. The 8mm one that you found would probably still work, just be aware that if it sticks through the header too much, it might contact the case.

 

So the length of the "mating pin" / Dimension "B" is actually 6.8mm?
I thought from original 3M datasheet it was 8.08mm / 0.318" freedom units ;) based on the codes:
Dimension A in the original 3M datasheet btw.
image.png.a964c7b96254662f2355a90ca59f1faf.png

image.thumb.png.db66b36d2bfa43b7352521ca1647181f.png

I have standard headers which are 6MM long on top btw.

 

Edited by Smithy
Dimension A in 3M datasheet not B
Link to comment
Share on other sites

The 3M headers seemed to be easy to source and typically had good stock. It is also less effort to keep track of one part series. The 6mm "standard" single-row header might work (see above post) but the extra fractions of a mm help. That is why the 8.08mm part was ordered. Feel free to stick with 8mm. The same TE part on Mouser/Digikey is discontinued.

Remember that the TE part you found has a plastic base height of 2.8mm, whereas the 3M one is "only" 2.54. So already there is a difference of 1mm.

Link to comment
Share on other sites

Just now, latigid on said:

The 3M headers seemed to be easy to source and typically had good stock. It is also less effort to keep track of one part series. The 6mm "standard" single-row header might work (see above post) but the extra fractions of a mm help. That is why the 8.08mm part was ordered. Feel free to stick with 8mm. The same TE part on Mouser/Digikey is discontinued.

Remember that the TE part you found has a plastic base height of 2.8mm, whereas the 3M one is "only" 2.54. So already there is a difference of 1mm.

Ah I get you now!
Thanks a lot for the help Andy!

Link to comment
Share on other sites

Ordered and installed longer 3M headers from farnell, they have the same height above the plastic where the the enter the shrouded header, but longer on the otherside but since this is cut at the joints it works out perfectly.

I'm at the stage where I'm testing the DIN switch function and SW17, SW18 and S20 on the left LE MEC 1.0 PCB are unresponsive in the MIOS terminal.
All other switches including SW19 and the Encoder switches work fine giving the correct debug messages.
All switches are pushed in and are not soldered yet.
Have them connected to the ENC-Plate 1.0 PCBs but these probably have nothing to do with my issue.
Pat_00 had a similar issue previously in the thread, but R3 seems okay to me:


Here's the view of the bottom of the LEMEC PCB:


IMG_20210524_172221__01.thumb.jpg.2db8d6d5fc56d0c80c085107af9d4369.jpg

Any ideas which components might be responsible?
Btw theres a lot of flux on the board from piggybacking the BAT54 diodes on the transistors. Its a no clean flux according to the Kester manual.
Thanks again!
 

 

 

Edited by Smithy
Fixed orientation of picture
Link to comment
Share on other sites

Just checked the Erata for that pcb on the wiki and it says:

The C3 cathode row is missing a PCB trace. Join the two legs of SW19 as shown in pink, preferably on the bottom of the PCB. The LED cathode leg could simply be bent across to the switch pin.
http://www.midibox.org/dokuwiki/doku.php?id=seqv4plus_le_mec

Will give this a go!

Edited by Smithy
Link to comment
Share on other sites

2021-05-24.thumb.PNG.5593e41c09a91ed1b8f629bb0aaf0981.PNG

The switches SW17/18/20 won't work in the matrix unless this is bridged, actually done by the switch SW19 itself. Try to insert a MEC switch here. Sometimes it will work without soldering, sometimes you need to solder it in (or otherwise bridge it) to get a signal.

Link to comment
Share on other sites

3 minutes ago, Smithy said:

Just checked the Erata for that pcb on the wiki and it says:

The C3 cathode row is missing a PCB trace. Join the two legs of SW19 as shown in pink, preferably on the bottom of the PCB. The LED cathode leg could simply be bent across to the switch pin.
http://www.midibox.org/dokuwiki/doku.php?id=seqv4plus_le_mec

Will give this a go!

Also a good catch, I had forgotten about that! There is not much risk in soldering in the MEC switches. It is only the Matias switches that you really want to be sure about because once they go in you have to desolder all of them to get to the inside of the PCB stacks (BC808s).

Link to comment
Share on other sites

10 minutes ago, latigid on said:

Also a good catch, I had forgotten about that! There is not much risk in soldering in the MEC switches. It is only the Matias switches that you really want to be sure about because once they go in you have to desolder all of them to get to the inside of the PCB stacks (BC808s).

Success! I inserted a pin from a header into the Cathode hole for the LED and used a crocodile clip to jumper the leg of the switch to it and all 3 switches work. :)
So in my case I didn't need to solder the switches first.
Going to insert the LEDs now and solder them in. Edit- actually the RJ and SJ resistors are next.

OLEDs are working also (even though thats later on in the guide).
I'm waiting on shielded MIDI sockets to arrive from Post NL so I won't be finished til I receive them but I'm pretty (Thorsten) klose to finishing it. ;)

Edited by Smithy
Link to comment
Share on other sites

Having some  further issues with the finished LeMEC 1.0 (left hand) board unfortunately.
When I load seq_l the Superflux LEDs of SW4, SW8, SW12, and SW16 light up and stay on.


IMG_20210525_152613.thumb.jpg.2db563755df83ba8635f02badbf28b27.jpg

All switches work fine including Switches SW 17 to 24.
But when I press any switch from SW17 - SW23, the LED of SW24 also lights up dimly when it shouldn't. The respective switches press light up fine.

I thought IC5 - the 74HC165 was the culprit as some pins were bent so I de-soldered it.
Unfortunately I broke pin 13 / pin C clean off  while trying to straighten it. The issue with the Superflux LEDs staying on when they shouldn't be existed before this btw!
 

image.thumb.png.e78a9af58995eeb968e3c1f144562649.png

I re-soldered the chip again hoping that it was unused, but now Encoder 6 is unresponsive when turning it clockwise / counterclockwise.
At least I presume pin 13 is responsible for this encoder, i could be wrong.

No big deal there, I can buy another 74HC165 off ebay and replace it.

Any ideas where the issue may be with the Superflux LEDs being lit up?
The LEMEC RH board works perfect, and this LEMEC 1.0 board has the same issue wether LEMEC RH is connected or not.

The fluxtest script works fine too, the LEDs that are lit up just turn yellow when the colour sequence passes.

Also the other leds turn red when pushing down on the econders and of course the naughty 4 change to yellow when they're pressed.
 

Edited by Smithy
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share

×
×
  • Create New...