EEZ Bench Box 3 (BB3)

[prasimix]

Good to see that you succeed. Do you know what the problem was, to possibly add it to the User manual, to avoid future frustrations of other users?

Do you still have a problem with Master: Failed at startup? The reason for this may be the wrong set date/time, the problem with the SD card or with the backplane. Additional info can be found in the event log (see Section 9.1).

[prasimix]

Hi,

I have a question regarding the usage of the BB3 (I am using Firmware V1.0 as shipped).
How do I change the default power up voltage/current (DCP405)...........seems to be 20Vdc/2.5A at the moment.

Please consult Chapter 7.

Also, and perhaps a future firmware tweak......I'd like to specify the minimum step change of the rotary encoder (say 0.1V, 0.01V etc), making it quicker to achieve the exact setting rerquired.

Already implemented . You can do that with User Sw set to "Encoder step" (see Section 9.8.)

[IanJ]

Hi all,

Based on the latest firmware (MCU) here are my initial thoughts of the BB3 GUI.

1. The encoder acceleration takes a bit of gettiong used to. Set it quite low and not much happens. Set it high and as you rotate the encoder with one quick action you might go up or down say 0.3V......but the acceleration only seems to come into play when you rotate the encoder again....i.e. it's that gap between rotating once and rotating it again that governs when the display should jump in much bigger steps. I am thinking there must be a better way for this.....it is really frustrating to use.

2. The encoder know is too big and too close to the User SW. I have lost count the number of times I have rotate the knob but accidentally hit the button whist doing so. I fitted my own, small diameter knob.

3. When using the encoder it's quite hard on the eyes to see the setpoint (Vdc) display change, i.e.:
25V
25.005
25.01
25.015
25.02
The readout jumps from side to side as it centre justifies all the time. It would be better to have the XX.XXX at all times. No digit blanking, and perhaps left justify maybe.
Personally, I would like to be able to only allow 0.1Vdc steps when using the encoder, but make this selectable in config for those that want max resolution.

4. Similarly, for the Vdc feedback, you can get the following on the display
8.1
8.015
8.005
8.00
or, indeed:
37.01
37.005
37
So, there's inconsistency here.......

5. The touchscreen buttons are probably a little on the small side, not much can be done about that I guess......but maybe a larger LCD next time.

6. Entering Voltages using the keypad, 40.1Vdc of course is not accepted, neither is 40.000002.......but 40.000001 is accepted?

So far though I am really liking the BB3.......hopefully it will become my primary PSU.

UPDATE: Just saw your post about USER SW.......fixes the XX.XXX issue for the most part, but would still like to see XX.X on the voltage setting for instance. at the moment it goes 14.8, 14.9 then 15. I think should be 15.0.

Ian.

[IanJ]

Good to see that you succeed. Do you know what the problem was, to possibly add it to the User manual, to avoid future frustrations of other users?

Do you still have a problem with Master: Failed at startup? The reason for this may be the wrong set date/time, the problem with the SD card or with the backplane. Additional info can be found in the event log (see Section 9.1).

I had to right click on STM32 BOOTLOADER in Device Manager and manually update the driver from list of available that popped up. DFU mode driver (various versions) was listed so I chose the newest.
It's possible because I already had an STM32 dev system loaded that it didn't install correctly.

Ian.

[prasimix]

prasimix,

With Keysight IO libraries I can query/response the BB3.....but from my own VB app I can make a valid connection but can't query the BB3 at all. I get an error responce.

I have written my own VB app but it uses an extensive GPIB library from PavelW. The IO library returns "MAV bit not set" when I try to query the BB3.
The developer of that library I remember he came across somebody else with the same issue try to connect to a DMM and offered the following repsonce at the time.

If the message is "MAV bit not set" then it is clear: your meter is not quite compliant with the 488.2 specification on the meaning of status byte bits. I explain. Polling the status byte is used here to check if the device is ready to send response, according to 488.2 this information is coded in the "Message Available" bit of the status byte, the bit 5 (so the default mask used is 16).
Keysight interactive IO works because it does not use any information from the status byte: when you hit "read" it calls the raw low-level read function which waits until it gets response.  But then once a "read" is issued the whole gpib bus will remain blocked until your meter is done with the measurement - very annoying if you have more than one devices connected, especially with high NPLC where it can take seconds to respond.  This software uses periodic polls to check if the data is available, before calling "read", in this way the bus remains available and the program is able to query two (or more) devices simultaneously.  This is the default configuration but it can be easily disabled on per-device basis (you may do it yourself or ask Ian to add checkboxes controlling the "enablepoll" flag for each device).  For a more elaborate version, the MAV masks can also be set individually for each device.

The above just my quick initial ideas as to the problem.........
I can work around it per what Pavel suggests and implement a checkbox in my app to enablepoll.........but just wondered initially your thoughts.

Ian.

Thanks for this. Martin will investigate and we'll let you know what he found. If you have a chance to get some VB sample, it could speed things up for me.

[IanJ]

prasimix,

With Keysight IO libraries I can query/response the BB3.....but from my own VB app I can make a valid connection but can't query the BB3 at all. I get an error responce.

I have written my own VB app but it uses an extensive GPIB library from PavelW. The IO library returns "MAV bit not set" when I try to query the BB3.
The developer of that library I remember he came across somebody else with the same issue try to connect to a DMM and offered the following repsonce at the time.

If the message is "MAV bit not set" then it is clear: your meter is not quite compliant with the 488.2 specification on the meaning of status byte bits. I explain. Polling the status byte is used here to check if the device is ready to send response, according to 488.2 this information is coded in the "Message Available" bit of the status byte, the bit 5 (so the default mask used is 16).
Keysight interactive IO works because it does not use any information from the status byte: when you hit "read" it calls the raw low-level read function which waits until it gets response.  But then once a "read" is issued the whole gpib bus will remain blocked until your meter is done with the measurement - very annoying if you have more than one devices connected, especially with high NPLC where it can take seconds to respond.  This software uses periodic polls to check if the data is available, before calling "read", in this way the bus remains available and the program is able to query two (or more) devices simultaneously.  This is the default configuration but it can be easily disabled on per-device basis (you may do it yourself or ask Ian to add checkboxes controlling the "enablepoll" flag for each device).  For a more elaborate version, the MAV masks can also be set individually for each device.

The above just my quick initial ideas as to the problem.........
I can work around it per what Pavel suggests and implement a checkbox in my app to enablepoll.........but just wondered initially your thoughts.

Ian.

Thanks for this. Martin will investigate and we'll let you know what he found. If you have a chance to get some VB sample, it could speed things up for me.

My code is based on Pawel's original here:-

https://www.codeproject.com/Articles/1166996/Multithreaded-communication-for-GPIB-Visa-Serial-i

I will be playing around with it this weekend to see if it is something else causing the problem as Pawel gave me that workaround which I haven't tried yet properly.

Ian.

[IanJ]

Good to see that you succeed. Do you know what the problem was, to possibly add it to the User manual, to avoid future frustrations of other users?

Do you still have a problem with Master: Failed at startup? The reason for this may be the wrong set date/time, the problem with the SD card or with the backplane. Additional info can be found in the event log (see Section 9.1).

Fixed......I didn't realize the SD card needed to be fitted.

Ian.

[prasimix]

Hi everyone,

@cadwal, @Kean: Kind of a solution for the power on settings: In the User Profiles Menu you have the option "Autorecall on power-on" (shown and described in the manual on page 51) which means it powers up with the settings you had the last time. Maybe another option that fits to your question: Settings->System protection settings->Force disabling of all outputs on power up.

Assembled my BB3 today. First impressions are pretty good so far. Here's what I noticed (positive first):
+ This thing is FAST. Boot time, UI response time and especially control and read out of the outputs. I've been sitting here for a minute turning the thing on-off-on-off just because it can do it so fast. The fast readout is impressive, too.
+ Manual looks well made. A lot of pictures of the UI with markings where to click what. Haven't used it a lot because IMO you shouldn't need a manual to get to know the basics about your device/tool/whatever. Manual is for the advanced and complicated stuff; the daily stuff should be self-explainig.
+ Many parts of the UI are well designed. Not only with regards to the previous point, but also with an eye on productivity. This was a big fear of mine; having a lot of features and a touch display is nice, but if it's slower to use than the good old 4 knob analog supply, it's annoying. Very happy to see that this fear did not become true for most parts.
+ Assembly was pretty straightforward. Big thumbs up for requiring only two types of screwdrivers, two types of screws (technically three if you count the two longer ones) and nothing else. No washers or clips or dunno what.
+ Packaging was pretty efficient. Everything protected and pretty much as compact as possible. This is the kind of details that give me the feeling that someone put effort and thoughts into it.
+ Putting those power supply modules into the bottom PCB was veery satisfying 
+ Look and feel during the construction was very good. There're only very few minor things that could be improved in my opinion. Like it!
+ Thaaank you for the option to turn off the beep sounds. Otherwise I'd have desoldered the thing sooner or later.
+ Not directly related to BB3 itself, but I can't emphasize enough how much I respect the effort prasimix put in the fulfillment of the crowd supply campaign. I really hope that in the end you didn't loose money on this; that's really the last thing you deserve.

Now on the other side I also noticed things that could be improved in my opinion, and how I think it could be done. This time I try to put it in a meaningful order. While I'm typing this I realise it's getting muuuuch longer than I thought. Sorry for that. I really want to give some constructive feedback with this and hope that it is indeed useful (and not a much too long waste of time..).
At the very end I describe a more serious issue with my DCP405 modules. So in case you want to read this review another day (which I fully understand), I'd appreciate if you could take a look at the very last point (made it bold).

My Setup:
BB3 with 2x DCP405 and 1x DCM220. I'm still on v1.0 (reasons below). I guess some of the issues have been fixed in the meantime.

Getting started:
- So you get this note with half a dozen links. They are short but still, you have to type them by hand. I think you could include a QR code. There're too many links to include a direct QR code for every link, so I'd suggest to include one single QR code. It'll link to this document online. Then you can copy&paste (or even better: click) all the links you need or want to have a look at. Of course, not everyone wants to do this on his phone, nor can everyone do this. But those people can still type the links by hand. There's no disadvantage for them.

Access to information:
- Index in the PDF manual does not link to the pages. That's usually one checkbox at the PDF export and very useful. Another thing that's a bit confusing/annoying: Since the index does not link to the pages I enter the page number in my PDF viewer, f.ex. page 23. But this actually brings me to page 25. There's a mismatch; the x-th page of the document is not page with number x.
- I understand that this device is quite complex and powerful. However I had to click quite some time around before I found the assembly instructions. I'd have to take a more serious look at this and especially think about how one could make this easier before I can come up with concrete ideas.
- At one point I wanted to download the manual, because I had no idea where I'd find this information on the website. I found a "User Manual" entry in the index table at the top (reminder: I'm on my smartphone). Two download buttons are presented to me, for english and german. Okay so I click on the download button for the german one. But instead of getting the PDF I get to a new page whose purpse is apparently to give me "Download Details". And another download button. So I click on that and what do I get? well, still no download. Another page, this time providing me the very useful "Download summary". Which is basically the same thing as the two pages before but at least a different layout. Oh, and there's another one of those wannabe download buttons. You'll never guess what it does. It actually started a download - finally!
Seriously, I doubt its necessary or a good idea to let anyone go through that many useless (IMO) pages to open a manual...
- Another thing I particularly noticed on the page about how to do a firmware update. There's a short intro and then a looooong Linux secion with loots of code. Might be nothing complicated and really well described - that's not the point. I am not a Linux user (I guess like most people) and this was all I saw, even when scrolling down a bit. It's only after quite some more scolling that the Windows tutorial appeared. What if I have a Mac? do I scroll even more? The point I want to make is: obviously this page contains a couple very long sections. they have to be that long and that's fine. But it does not make the page more readable. If I have to scroll through an entire webpage that is multiple screen heights long, to find out if it even contains the information I'm looking for ("Firmware update with Windows"), I think the design is not the most user friendly.
Here's what I'd suggest for the mobile webpage. The page filling "global index" at the top can be hidden in a side menu. You know, the thing you acces with the button in the top left corner, that has thee horizontal lines as icon. This is important because the first thing I will see when opening the page can now be actual content and not a menu have to scroll through every time I click on a link. For the page itself I think a wikipedia-like design would help a lot. What I mean is that every section is at first collapsed and only its title visible. This is IMO a very clever way to allow quick navigation through a page with long sections - like wikipedia articles or firmware update guides. I'd immediately see 13.1 Linux, 13.2 Windows - perfect, tip to show that section and done.
- Speaking of firmware updates (does not quite match this section but fits perfectly after the previous point). You need different tools to upgrade different parts of the BB3? That's a bit disappointing... I hope you manage to get this all under one hood. Another issue IMO: you need to get this wired up to a PC...? Not sure if everyone has a lab PC - I for sure didn't have one for a loong time. What I'd love to do is this: Download the latest firmwares, drag them onto the microSD card, and the BB3 will update everything it finds a firmware for. No need to have a PC next to your lab supply, no need to download and configure any drivers or software according to multiple (!) tutorials. Just drag&drop, plug&play, done. (An even more fancy idea: since the thing has networking... You could let BB3 download new firmwares from the github release page  )
- The complicated looking update procedure is the main reason why I'm still on v1.0. Download, drag&drop would've been a no-brainer but this is something I have to read through on another day. And while I absolutely appreciate that the firmware can be updated (meaning the device gets better and better over time), this makes me think that not as many people will update their BB3 and they won't do it as frequently as they would otherwise. HOWEVER, as I said, I haven't gone through the procedure, maybe it's already very quick 'n easy and half of my critic here is not right. Would be great! Still want that SD update 

Case/Mechanical/Assembly:
- I did not count the screws at the beginning but I know I missed one black screw when I was done. For some reason I have 6 of the silver screws left. Haven't found a hole where they'd be missing... That aside, I do think there should be 1 spare screw of both types, possibly even a third of the longer black ones. Since there're only 2 (3) different screw types, this shouldn't add too much cost and would help in case you loose one. Especially since quite a few BB3 users will open their case again to upgrade/change the modules - another chance to loose a screw. I thought about this when I realized there were no extra screws. And I really wished it when I realized I missed one. Don't want to blame anyone but I assembled it on a clean table with a clean, white floor underneath and I can not find it. So yeah, I'm probably blind. 
- The screws that hold the display are rusted...? Is that normal?
- About half of the threads in the case had paint in them which made it pretty hard to force the screw in. At least they won't get lose! Don't know if this can be avoided or not; just wanted to note it.
- The top case is slightly bended (higher in the middle); it can be pushed down about 2mm. Additionally, because it's screwed down on only two axes (left and right on the same place), the top part can be slightly pushed forwards/backwards and doesn't line up with the bottom part. Both things are purely cosmetic and no real issues but IMO this does not match the quality and feeling of the other parts. Take the bottom half; that one feels rock solid. Second reason for brining it up: both things could be fixed with one minor change I think: add 1 screw between front panel and top case, and 1 between back panel and top case. This should not only prevent the forward/backward fexibility but also hold the top piece down properly for an overall even more rigid feeling.
- The cable for the fan is about 3cm too short? I can't put it the way as shown in the assembly manual. If I put it from the connector down between the DC/DC converters and the heatsink, there's not enough length to put it against the backplate; it goes straight to the fan. (think of a triangle. hypotenuse is shorter than the other two sides together). Is this meant to be like this? Again, doesn't cause any issues I think but didn't seem right to me...
- A true detail: Every word on the front panel and on the module front panels has the first letter in upper case - as I'd expect it. Except of two things: SD card, which is fine IMO, and "Open hardware / open source". The second open in lower case does not look right to me; doesn't match the first one. You dont write "Dual sync buck" instead of "Dual Sync Buck" either. Not 100% sure if the "hardware" and "source" should be upper case, too, but at least the two "open" must be equal IMO. As I said, a detail, but one I'll see every single time I use the BB3 - which will hopefully be very often

The Rotay Encoder
Now this one deserves its own section in my post because it was honestly the biggest disappointment. From your videos I can see that you don't use the encoder that much and prefer the keypad. I tried both and the encoder is really the thing that suits me the best. I'm sure this is very individual and every one will come up with a different habit that seems perfect to him/her. The most natural feeling way for me is touch the value, turn the knob. I don't use the keypad, I don't use the knob's push button to cycle through the values. What made it even more disappointing is that I absolutely believe that it could be possible to set values very fast. Like "I don't miss the 4 potentiometers from my analog supply at all" kind of fast. So here we go with my issues...
- For some reason the encoder does not sit straight. This means that the knob is too close to the front panel in one point and depending on how I turn it, scratches over the front panel. This feels really really cheap. I might even open it again and try to resolder the encoder to fix this.
- The knob sits rather loose on the encoder shaft. Being a big know makes it look even worse. Again, very cheap feeling to me.
- The encoder itself has no decent ticks but feels rather soft and spongy - again, a big knob doesn't help. Big knob = large radius = large lever arm => soft spongy ticks feel even more soft.
- The velocity control does not work at all for me. So I did play quite a bit with it and here's the issue (I think): the velocity increases only after about half a turn. If you turn the knob a quarter turn you won't get very far no matter how fast you do it. However, if I turn the knob with two finger, half a turn is the most I can turn it; usually more like 1/3 of a turn. If the knob was smaller, it would be more; maybe someone with bigger hands can turn this big knob a full turn but I cant. Same thing if you turn it with one finger by putting it on the side; you get about 1/3 of a turn.
So for my usual way(s) of using a knob velocity control is de facto inactive, making the knob de facto useless.
The only way to really get velocity control is to put one finger on the front of the knob and turn around (which includes slipping off). Then you can turn it a full revolution (before slipping of) and velocity control becomes veeery visible. But as described above; for the way I use a knob this is much too late.
Bottom line: please adjust the velocity control such that you can easily get across the range with 1/3 turns. I know it's possible. And I'm pretty sure some knobs I've used recognize when you to multiple quick movements (like you usually do with two fingers), and don't restart for every 1/3 turn at "zero velocity". One might say they have a bit more inertia.
- Even worse: If I have the velocity settings not at maximum (gets worse the lower you go), I can turn the knob and individual steps don't do anything anymore. I can turn the knob a whole revolution without changing the value. I think 1 tick should always do something.

Using BB3 and its UI (reminder: v1.0!)
- While I said above that I do think it's well made in most parts, it does make a ... convoluted? impression. Even though I know what is what, the overview with the 4 channels (I have 2x DCP405 + 1x DCM220) does somehow not look "clear", "stuctured". I honestly can't tell you why.
- When I tested it for the first time I turned on Channel 1 (first DCP405); fine. Tipped on channel 2 (second DCP405) - and I get a warning that its not calibrated. Okay, no worries, let's do it - but nooo I can't. This thing comes out of the box with a cal password. Okay, it's nice to have that option, but the few people who need this, will find the option because button for changing the cal password sits right next to the button for the main password - no chance to miss it. But everyone else (and seriously, how many people really want this?) has additional work with this. And especially during the very first "hands on", I shouldn't have to search through the manual for some default password... Even worse - I can't remove the thing? I can't leave the password blank when changing it? Might someone explain, why I, as a private user, would need to be forced to password protect the calibration routine?
- This point I'll bet that it's been fixed by now: There's the Settings point "Encoder & User sw" but I don't see any settings for the user sw? Maybe I'd have to read through the manual in detail but so far I haven't found any option or setting for that switch.
- I kinda mentionned it in the positive points but that beep thingy sounds "not good"; rather cheapish-annoying. Might be highly subjective; maybe other users like it.
- Now this is serious IMO. Using the DCP405 modules (happens with both of them). When I have the current limit at 8mA or lower and I short the output, it does not recover correctly after removing the short. Voltage was around 10V, but doesn't seem to matter. What happens is that the voltage stays slightly below the set value (f.ex. 11.5V instead of 12V) and the CC LED remains on. Current stays at the limit until I increase it to about 10mA or higher. Only then the current will drop back to 0 and I can decrease the limit again. MAYBE this a soldering fault on both of my modules...? One of them hat 2 capacitors that fell of the board, and a third one with a bad solder joint. Additionally I just noticed that that module must have at least one more bad solder joint, because the CC LED started flickering, and reacts if I knock on the case. Any ideas? That aside, the issue that it does not recover after entering current limiting mode is equally present on the other DCP405 module. I couldn't find any bad solder joints on that one but maybe I missed something?


Okay, finally, I'm done (or forgot the remaining points). If you read through all of this, I do hope it's useful. Thank you for all of your work, and in case all of this sounds too negative, you should know that I don't regret to have opted for the BB3 and would do it again. 

Max

Massive thanks Max, for massive feedback. That is what we need. Seems to me that you are a perfect beta tester! I would love to get a similar review in a timely manner in the future.

I don’t know if it makes sense to answer every point so I would make some summary.
The complete enclosure design is a big disappointment considering that the same manufacturer made the enclosure for the previous project (H24005) and this was supposed to be a simple upgrade. Unfortunately that did not happen. I don’t know if it makes sense to say how much time and money was spent and prototypes made to end up getting something so mediocre.
You yourself have already noticed some things, such as rusty studs for the display frame because an aggressive cleaner was used in preparation for painting. I’m sorry I didn’t have the opportunity to deliver something better.

Your comments about the website are accurate. I can't say anything other than that the website requires a complete redesign that will, among other things, take better care of mobile browsers. This includes the download procedure, for which we use the Joomla extension in which everything is so complicated.
We will probably have to hire an external company for that, because we cannot do it alone at the moment.

The things we will pay attention to is definitely the encoder whose support is weak and not in line with everything else that has been done so far. The second thing we will work on is definitely further simplification of the system (master) firmware download procedure.

I have already answered the User Sw menu question, but I repeat here: to enter, press and hold User sw until the menu appears (see Section 9.8).

In the end, I can say that I cannot repeat the behavior of DCP405 that you mention. To reproduce the problem I will need more details about the procedure, maybe some short video if it is not a problem?

[prasimix]

Hi all,

Based on the latest firmware (MCU) here are my initial thoughts of the BB3 GUI.

1. The encoder acceleration takes a bit of gettiong used to. Set it quite low and not much happens. Set it high and as you rotate the encoder with one quick action you might go up or down say 0.3V......but the acceleration only seems to come into play when you rotate the encoder again....i.e. it's that gap between rotating once and rotating it again that governs when the display should jump in much bigger steps. I am thinking there must be a better way for this.....it is really frustrating to use.

2. The encoder know is too big and too close to the User SW. I have lost count the number of times I have rotate the knob but accidentally hit the button whist doing so. I fitted my own, small diameter knob.

3. When using the encoder it's quite hard on the eyes to see the setpoint (Vdc) display change, i.e.:
25V
25.005
25.01
25.015
25.02
The readout jumps from side to side as it centre justifies all the time. It would be better to have the XX.XXX at all times. No digit blanking, and perhaps left justify maybe.
Personally, I would like to be able to only allow 0.1Vdc steps when using the encoder, but make this selectable in config for those that want max resolution.

4. Similarly, for the Vdc feedback, you can get the following on the display
8.1
8.015
8.005
8.00
or, indeed:
37.01
37.005
37
So, there's inconsistency here.......

5. The touchscreen buttons are probably a little on the small side, not much can be done about that I guess......but maybe nding problem, obviously something is not done well because there is no reason to accept 40.000001 and not accept 40.000002. It must reject both. Reported as issue [ua larger LCD next time.

6. Entering Voltages using the keypad, 40.1Vdc of course is not accepted, neither is 40.000002.......but 40.000001 is accepted?

So far though I am really liking the BB3.......hopefully it will become my primary PSU.

UPDATE: Just saw your post about USER SW.......fixes the XX.XXX issue for the most part, but would still like to see XX.X on the voltage setting for instance. at the moment it goes 14.8, 14.9 then 15. I think should be 15.0.

Ian.

As I already answered @MaxZ encoder functionality needs to be improved. Not much has been done on it because I don’t use it much, but that’s by no means an excuse. I opened a new issue for further discussion and tracking (#72).

Thanks for issue #6! I like such details. Although this is a rounding problem, obviously something is not done well because there is no reason to accept 40.000001 and not accept 40.000002. It must reject both. Reported as issue #73.

[prasimix]

- The cable for the fan is about 3cm too short? I can't put it the way as shown in the assembly manual. If I put it from the connector down between the DC/DC converters and the heatsink, there's not enough length to put it against the backplate; it goes straight to the fan. (think of a triangle. hypotenuse is shorter than the other two sides together). Is this meant to be like this? Again, doesn't cause any issues I think but didn't seem right to me...

I forgot to comment on this. My fan cable looks like this:



We just added a Molex connector to the original cable. Do you think it should be even longer? Extending the cable would be too complicated.

[MaxZ]

Hi prasimix,


Thanks for the response to my feedback! If it helps I‘m more than willing to test new features if I find the time to do so.
I guess in that case you‘d prefer another way of communication than this forum thread...? Especially since there‘re now 200 people with issues that may post here.

Kinda good to hear that you could not reproduce the issue with the DCP405; this probably means it‘s not a general design flaw and rather an issue with my specific modules.
here‘s the video of the DCP405 modules which don‘t recover. It unvoluntarily features the knob again... you‘ll see what I mean. Btw I could not reproduce the flickering CC LED that seemed to be caused by a bad solder joint.
https://youtu.be/QlECKDh7nF0

Since I was already recording I thought I‘d record my issues with the rotary knob, too.
https://youtu.be/oOLDDVv1HUU

Considering the fan cable: I attached pictures of my cable when trying to put it down as seen in the instructions. It‘s just long enough to get down to the bottom. If it was possible I‘d lile to see a little more length there because I feel like the other cables (whose leghts are good btw!) could possibly push against this one, putting it under stress. But seriously, it may be just fine. Since an extension of the cable is indeed a lot of work, it might be possible to „mirror“ the PCB, putting the connectors at the bottom instead of the top? That would give the few centimeters I‘m looking for. This however comes at the expense of worse accessibility of the connectors... It‘s something I‘d have to spend a bit more time on, maybe try it out. Hard to judge without opening it again.

Edit: Sorry for them being upside down... hate jpeg auto/not-auto rotate issues.


Kind regards,
Max

[prasimix]

Hi prasimix,


Thanks for the response to my feedback! If it helps I‘m more than willing to test new features if I find the time to do so.
I guess in that case you‘d prefer another way of communication than this forum thread...? Especially since there‘re now 200 people with issues that may post here.

Right, it should be a different arrangement and you should have on disposal new things that need to be examined.

Kinda good to hear that you could not reproduce the issue with the DCP405; this probably means it‘s not a general design flaw and rather an issue with my specific modules.
here‘s the video of the DCP405 modules which don‘t recover. It unvoluntarily features the knob again... you‘ll see what I mean. Btw I could not reproduce the flickering CC LED that seemed to be caused by a bad solder joint.
https://youtu.be/QlECKDh7nF0

This is definitely not a normal behavior. Even with down-programmer disabled it should go down to zero fast. Ok, I'm testing with the latest firmware but that should not make a difference. That issue requires more examination.

Since I was already recording I thought I‘d record my issues with the rotary knob, too.
https://youtu.be/oOLDDVv1HUU

I'm sharing your frustration about encoder (Bourns) + knob, and I’ve been living with it for a few years now! The fact is that I did not find a satisfactory combination of knob and encoder. Tolerance is unacceptable, almost 1 mm. I remembered my Denon receiver that I paid a small fortune few years ago. I can tell you the tolerance of his knob is no better, but that doesn’t comfort me. If you or someone else knows about quality encoders and knobs, I would love to try a new combination.

Since an extension of the cable is indeed a lot of work, it might be possible to „mirror“ the PCB, putting the connectors at the bottom instead of the top? That would give the few centimeters I‘m looking for.

It's not that simple because on the AUX PS we have mains voltage and traces towards the power switch go along the bottom edge. I can't put a fan connector down there.

[IanJ]

As I already answered @MaxZ encoder functionality needs to be improved. Not much has been done on it because I don’t use it much, but that’s by no means an excuse. I opened a new issue for further discussion and tracking (#72).

Thanks for issue #6! I like such details. Although this is a rounding problem, obviously something is not done well because there is no reason to accept 40.000001 and not accept 40.000002. It must reject both. Reported as issue #73.

Just to add for completeness on the encoder issues:-
1. There seems to be a change of direction issue with the encoder....i.e. if turning up the way and you decide to turn it down it takes an extra couple clicks on the encoder to reach the same value.
2. I have set the encoder steps to 100mV and so would expect to have one click per 100mV.....but some clicks don't change anything.

Other than that, could you add 0.5V for a valid step change. I for one would find that useful as 100mV maybe too small, 1V too big.

Ian.

[prasimix]

Thanks, I updated #72 with your latest comments, and yes, we can add 0.5 V step. You can expect some progress about encoder issues very soon!

[IanJ]

I'm sharing your frustration about encoder (Bourns) + knob, and I’ve been living with it for a few years now! The fact is that I did not find a satisfactory combination of knob and encoder. Tolerance is unacceptable, almost 1 mm. I remembered my Denon receiver that I paid a small fortune few years ago. I can tell you the tolerance of his knob is no better, but that doesn’t comfort me. If you or someone else knows about quality encoders and knobs, I would love to try a new combination.

I designed a product a couple of years ago which used a bourns 24PPR  encoder......and had exact same frustrating 'slap' issue.
Bourns PEC12R-4222F-S0024
Bourns PEC11R-4015F-S0024
Some of them have plastic shafts, others have metal ones.......if I remember correctly the metal shafted ones had a little less slap and were a drop in replacement. Might be worth looking at.

Ian.

[MaxZ]

Hello again,

Kinda good to hear that you could not reproduce the issue with the DCP405; this probably means it‘s not a general design flaw and rather an issue with my specific modules.
here‘s the video of the DCP405 modules which don‘t recover. It unvoluntarily features the knob again... you‘ll see what I mean. Btw I could not reproduce the flickering CC LED that seemed to be caused by a bad solder joint.
https://youtu.be/QlECKDh7nF0

Sorry, I forgot to mention that today I updated* to v1.3.2. So the video was made with v1.3.2 and the issue is still there, exactly the way it was with v1.0.
Updating was thankfully easier than I feared. Since I don't want to put another tldr post here, I'll save that (and other) feedback for another day and place 

Since an extension of the cable is indeed a lot of work, it might be possible to „mirror“ the PCB, putting the connectors at the bottom instead of the top? That would give the few centimeters I‘m looking for.

It's not that simple because on the AUX PS we have mains voltage and traces towards the power switch go along the bottom edge. I can't put a fan connector down there.

I am aware of this (or misunderstand the issue you describe). Hence I suggested to mirror the whole PCB instead of just moving the fan connector. So not only would the fan connector go to the bottom, but the mains voltage traces would at the same time go to the top. I know that this might not be possible because components can't be mirrored and not every component is symmetrical like a resistor. It was an idea, no more. And as I said, I'd need to have a look at more details of the layout of the PCB and the case to judge whether it is feasible or not. But honestly, I didn't want to dig that deep into it for now. Especially since I'm not even sure the cable is truly too short, which is the very first thing I'd check before thinking about any kind of redesign. The next time I open it up (which will probably be soon because of the DCP405 issues), I'll have another detailed look at it and give more precise feedback.


Kind regards,
Max

[prasimix]

I'm sharing your frustration about encoder (Bourns) + knob, and I’ve been living with it for a few years now! The fact is that I did not find a satisfactory combination of knob and encoder. Tolerance is unacceptable, almost 1 mm. I remembered my Denon receiver that I paid a small fortune few years ago. I can tell you the tolerance of his knob is no better, but that doesn’t comfort me. If you or someone else knows about quality encoders and knobs, I would love to try a new combination.

I designed a product a couple of years ago which used a bourns 24PPR  encoder......and had exact same frustrating 'slap' issue.
Bourns PEC12R-4222F-S0024
Bourns PEC11R-4015F-S0024
Some of them have plastic shafts, others have metal ones.......if I remember correctly the metal shafted ones had a little less slap and were a drop in replacement. Might be worth looking at.

Ian.

The choice of encoder is further narrowed because an angled version is required. What you sent are "straight" versions.

[prasimix]

I am aware of this (or misunderstand the issue you describe). Hence I suggested to mirror the whole PCB instead of just moving the fan connector. So not only would the fan connector go to the bottom, but the mains voltage traces would at the same time go to the top. I know that this might not be possible because components can't be mirrored and not every component is symmetrical like a resistor.

A complete mirror would also require moving down the IDC 16-pin connector which would complicate the connection to the MCU module.

[MaxZ]

Opened it up again, the fan cable is fine as long as you do not put it inbetween the heatsink and the DC/DC module as shown in the assembly instructions. If you do that, it'll not quite reach to the case, meaning it will stay under stress.
Thinking about the DCP405 issue... Since both modules show the exact same issue, is it possible that the actual problem is on another board, affecting both modules at the same time? Thought this would be more likely than having two modules fail in the exact same way...

EDIT: in a couple days I should have access to my IR camera again so maybe I‘ll find out where the power is dissipated.


Kind regards,
Max

[cadwal]

A BB3 arrived here in Sweden today. Dropped off at my door just after lunch today having left Texas on the 14th.

Assembly now started 

Oops, failed on step 2. Missing one J, have 5 or so extra L ...

So I left out the middle stud on the AUX-PS board as that seemed to be the one least necessary and easiest to add later.

But this time I failed on step 9. Nothing happens at all when I turn on the (nice) switch. Nothing on the screen at all and complete silence.

The fuses in the IEC connector are ok, cable F has the correct orientation and the flat flex cable also is as deep in as it can go.

[prasimix]

A BB3 arrived here in Sweden today. Dropped off at my door just after lunch today having left Texas on the 14th.

Assembly now started 

Oops, failed on step 2. Missing one J, have 5 or so extra L ...

So I left out the middle stud on the AUX-PS board as that seemed to be the one least necessary and easiest to add later.

But this time I failed on step 9. Nothing happens at all when I turn on the (nice) switch. Nothing on the screen at all and complete silence.

The fuses in the IEC connector are ok, cable F has the correct orientation and the flat flex cable also is as deep in as it can go.

Remove 16-pin IDC cable and check if standby LED is active when turn power on. Few other backers reports that cable failure. Also you should hear fan is spinning for a few seconds after the power is turned on.

[prasimix]

Opened it up again, the fan cable is fine as long as you do not put it inbetween the heatsink and the DC/DC module as shown in the assembly instructions. If you do that, it'll not quite reach to the case, meaning it will stay under stress.
Thinking about the DCP405 issue... Since both modules show the exact same issue, is it possible that the actual problem is on another board, affecting both modules at the same time? Thought this would be more likely than having two modules fail in the exact same way...

EDIT: in a couple days I should have access to my IR camera again so maybe I‘ll find out where the power is dissipated.


Kind regards,
Max

I would rather suspect the DCP405 modules themselves. Can you check if the same problem exists when the current range is set to High (see Section 11.5.1)?

[MaxZ]

I would rather suspect the DCP405 modules themselves. Can you check if the same problem exists when the current range is set to High (see Section 11.5.1)?

No, can't get the issue when setting it to high range. Going down in the 10s of uA range makes the CC LED pulse slowly. If it wasn't a fault it would actually be a nice "heartbeat effect":
The limit values were only reduced to get a better visibility on the graph. They don't seem to affect the behavior.
Maybe this slowed down version helps diagnosing the behavior. If I can provide any other information, please tell me!

Kind regards,
Max

[cadwal]

<snip>

So I left out the middle stud on the AUX-PS board as that seemed to be the one least necessary and easiest to add later.

But this time I failed on step 9. Nothing happens at all when I turn on the (nice) switch. Nothing on the screen at all and complete silence.

The fuses in the IEC connector are ok, cable F has the correct orientation and the flat flex cable also is as deep in as it can go.

Remove 16-pin IDC cable and check if standby LED is active when turn power on. Few other backers reports that cable failure. Also you should hear fan is spinning for a few seconds after the power is turned on.

Ah, yes, that gives fanspin and a green standby LED. Thanks!

I had removed the AUX-PS board and was checking it out and I could not find anything that was wrong (without connecting AC power that is), even the power button checked out ok.

I will bring out suitable probes and check that cable in detail...

[cadwal]

<snip>

<snip>

Remove 16-pin IDC cable and check if standby LED is active when turn power on. Few other backers reports that cable failure. Also you should hear fan is spinning for a few seconds after the power is turned on.

Ah, yes, that gives fanspin and a green standby LED. Thanks!

I had removed the AUX-PS board and was checking it out and I could not find anything that was wrong (without connecting AC power that is), even the power button checked out ok.

I will bring out suitable probes and check that cable in detail...

Ah yes, all pins are connected to all other pins, probably not what was intended 

The attached image is a bit fuzzy, but it can be seen that the cables are not in their "grove" but rather right in between so the "knifes" or whatever one wants to call them cuts through the cables on either side of it so everything gets connected together. Quite an achievement that...

If this was a TME part it probably is not the best place to get a replacement?