EEZ Bench Box 3 (BB3)

[Rerouter]

48V is because the same slot has to support the higher voltage single channel modules.

[prasimix]

48V is because the same slot has to support the higher voltage single channel modules.

Exactly, that will run cost down, and with unified DCbus, we can in the next step starts to think about AC/DC converter with multiple isolated outputs, and after that about bidirectional one when 2Q and 4Q power modules come into picture!

[prasimix]

A small update regarding the DCM220 V_in hack (see #144): I found that when LT3763 is disabled full DC_bus voltage appears on that pin (that should be input)! Therefore the voltage difference between added regulator MOSFET source and gate is 36 V, that is not good. To fix that issue a diode in series with MOSFET source to LT3763's V_in is added.

[prasimix]

The latest version of DCP405, a single output power module, is more or less finished and ready for planned forthcoming crowdfunding. It includes a few modifications (thanks to my friend Macola) from the previous revisions that were used the same design as the power module from the EEZ H24005 project. First, how it looks like:



Here is the list of changes:

• The main change is that N-channel MOSFET as pass element in the post-regulator section is replaced with 2 x PNP BJTs (D45H11).
• Changing of pass element driver section also asked for different type of down-programming. Therefore that circuit is completely redesigned.
• The CV/CC mode of operation detection is now much simpler.
• The bias power supply that was use not so efficient combination of linear step-down regulator (TL783) and LTC3260 capacitive switcher is replaced with low power buck (TPS54060) that is using coupled inductor to generate positive and negative voltage.
• The good thing about replacing MOSFET (that is in TO247 package) and TL783 is that now I can use "off-the-shelf" L-profile heatsink, that will greatly simplify assembling, and no custom heatsink is needed anymore.
• More precise (and still very cost attractive!) MC33272ADR2G op-amps are used instead of TL072. The CV loop compensation network is also cleaned up and adjusted to work properly with new op-amps.

Schematics of the post-regulator is shown below (mentioned new bias supply and CV/CC detection is not shown), and complete design is available on the GitHub as usual.



LTspice simulation of the new post-regulator is also available on the GitHub, but it is also attached to this post. Please note that it assume that input voltage is fixed, not regulated by pre-regulator circuit.

[prasimix]

Oh yes, I forgot to mention that I've made some measurements for the DCP405 module that can be found on: https://www.envox.hr/eez/eez-bench-box-3/eez-bb3-measurements/dcp405-power-module.html

[prasimix]

The firmware is now pretty completed regarding the new GUI, and it can be test even without having physical device thanks to the Web simulator. It includes almost all functionality of the real device and gives even possibility to attach load, and mouse wheel can be used to simulate incremental encoder. I'm attaching a few pages below and your feedback about look and feel, usability, etc. is highly appreciated.

[exe]

Wow, this is great! I used it a little bit, the confusing part that by default max current was set to zero. So, it was in CC mode, but voltage and current were 0. Sort of what it should, but in real world zero is not exactly zero, so there are some numbers jumping around.

I quickly used it, was a bit difficuilt in the beginning. There are many options. I'd prefer some dedicated knobs, but this design with several modules just doesn't support it. Nonetheless, I hope it would be possible to bind the rotary knob to quickly change voltage. Motivation: sometimes I characterize elements like diodes by slowly manually changing current/voltage and observe how it reacts. Sort of poor man's SMU. I understand that it supports programming, but I found a dedicated controls to be useful.

Anyway, great job!

[prasimix]

Yes, it's possible to "dedicate" encoder knob to any of output values (U, I or P). On the simulator you just need to position mouse on it, click, and use mouse middle wheel to increase/decrease current value. On the real device you just need to select output value with your finger thanks to touchscreen and engage knob. If you click "user sw" you can change knob step value (e.g. auto, 500 mV, 1 V, 2 V or 5 V when voltage is changing).

[mrpackethead]

cant' wait to buy one!

[prasimix]

Don't know if you possibly have the previous H24005 unit, but if you have one I can say that BB3's bigger display, better rendering (fonts are now antialiased) and responsiveness makes a huge difference, probably similar to one experienced between H24005 and regular PSU with LED/LCD displays, keypad/knobs. You cannot feel that on the simulator.

[AlanS]

Yep. I thought the previous simulator was great and then realised that using the real thing would mean I was better using a rubber topped pencil than my stubby fingers. STILL it was better then my other supplies.

[prasimix]

For those who are not aware that there is a topic for the upcoming crowdfunding campaign, here is the link:
https://www.eevblog.com/forum/crowd-funded-projects/eez-bench-box-3-sequel-to-eez-h24005/msg2786804/#msg2786804
The campaign should start this week and I want to thank you in advance for your support and I believe BB3 will live up to your expectations.

[jbb]

The campaign is now up.

I hope that I’ll be able to work with the DIB spec to produce custom modules, and I am looking forward to (hypothetical) extra modules for things like 2Q supplies and precision power meters.

I have a couple of technical questions:

• How are you doing on testing and assembly? I’ve found with my products that there can be quite a lot of design for testing work...
  
  Eg, the connectors that were fine for dev work may wear out surprisingly fast when you’re testing a production batch and cause intermittent test failures :-(
  
  

[li]If you get a good number of orders and the cost per unit comes down a bit, would you consider doing an IEC 60950 (or similar) safety evaluation in a test lab?[/li][/list]

[prasimix]

I hope that I’ll be able to work with the DIB spec to produce custom modules, and I am looking forward to (hypothetical) extra modules for things like 2Q supplies and precision power meters.

Many people are asking for 2Q module, and I wonder what is the best way to gather information on what people need and create an initial specification that should be used as a design guide. Perhaps we could start a discussion here and then create a new public repo on GitHub and transfer useful information there to make it accessible in a more structured way.

I have a couple of technical questions:

• How are you doing on testing and assembly? I’ve found with my products that there can be quite a lot of design for testing work... Eg, the connectors that were fine for dev work may wear out surprisingly fast when you’re testing a production batch and cause intermittent test failures :-(
  
• If you get a good number of orders and the cost per unit comes down a bit, would you consider doing an IEC 60950 (or similar) safety evaluation in a test lab?

Basic PCB testing and module assembly will be done by a PCBA contractor, functional testing will be done on my side when firmware will be downloaded and the functionality of the entire system will be tested. The self-test itself gives a good picture of what's is good or wrong, this was the case with the H24005 testing so I'll do the same for BB3.
I'm not in a position to invest in any testing automation yet, but I believe this will change if the number of units increases over time.
Having in-house safety (and EMC) evaluation could be a great plus, but it's also out of my current budget. So far, I've secured funding for entering EU-accredited lab for EMC and safety CE compliance certification for the current BB3 configuration. I'll see when testing starts if I can arrange with them some pre-compliance testing for future modules/configurations with some discounts.

[Rerouter]

for the 2Q, If I recall correctly it kind of needed your fancier mains supply, as you where planning on dumping the excess power back to mains, I suppose it would not be hard to take the current thermal power budget to know how much you could dissipate using the existing modules.

[jbb]

...
Many people are asking for 2Q module, and I wonder what is the best way to gather information on what people need and create an initial specification that should be used as a design guide.
...

I'll bet a quality 2Q module turns out to be quite complicated... for example, working out the voltage and current loop priorities in a sensible fashion.

...
Having in-house safety (and EMC) evaluation could be a great plus, but it's also out of my current budget. So far, I've secured funding for entering EU-accredited lab for EMC and safety CE compliance certification for the current BB3 configuration. I'll see when testing starts if I can arrange with them some pre-compliance testing for future modules/configurations with some discounts.

Yeah, gear gets expensive real fast.  And EMC testing sometimes requires large and specific setups (Faraday cage room, anyone?).  You probably know this already, but it's good to get in for compliance testing quite early - issues get more expensive to fix as you get further through the design process.  This might apply double for safety testing...

for the 2Q, If I recall correctly it kind of needed your fancier mains supply, as you where planning on dumping the excess power back to mains, I suppose it would not be hard to take the current thermal power budget to know how much you could dissipate using the existing modules.

While it might seem attractive to dump power to the mains side, you could get an islanding problem, and may require quite elaborate anti-islanding systems.  I think you're right that a good start would be to rate the unit to suit available power dissipation.

I guess you could have an external power resistor which plugs in for expanded range.  Getting a bit complicated at that point...

[prasimix]

It really depends. Lets start with presumption that 2Q should substitute electronic load. Electronic loads comes in all shapes and sizes depending of the use case. My idea is to start with "dissipative" 2Q module. If we want to dissipate e.g. 100 W or even more its obvious that such dissipation will be an overkill if everything has to be done inside BB3 chassis. If the main scenario is to test another power supply, DC-DC module or huge battery discharge then we cannot do that in that way.

But, if we'd like for start to address people how is willing to test their battery operated or low power devices (e.g. various IoT gadgets) that is operated with small battery and low voltage then it seems that 2Q with e.g. 20-25 W of sinking power could be more then enough, or not?

In the next step for bigger power e.g. up to 200 W we could provide connection for external power resistor, that has to be safely caged against the accidental touch, and dissipate excess power outside BB3 chassis. Not so elegant but could be cost effective.

After this stage all others should include bidirectional AC/DC power stage.

[prasimix]

Hi jbb, it seems that we're replying in the same time. I'm agree, making a quality (and reliable) 2Q could be quite complicated. Another mission impossible for me, but with assistance from here and there I'm now quite confident that it is doable.

[excitedbox]

Why not make the power supply battery powered and use the big battery to absorb the power instead of dissipating it. Or use super capacitors to store the power. If you keep an almost empty battery in the supply under normal conditions and only charge it that way and if it is charged beyond a certain point power other channels with the stored energy. Then instead of dissipating the power you store it.

[prasimix]

Why not make the power supply battery powered and use the big battery to absorb the power instead of dissipating it. Or use super capacitors to store the power. If you keep an almost empty battery in the supply under normal conditions and only charge it that way and if it is charged beyond a certain point power other channels with the stored energy. Then instead of dissipating the power you store it.

Hm, sounds interesting, perhaps supercap could be more suitable for such job since it should offer more freedom for charging i.e. when sinking is active. In that case a possible option could be a connector for connecting external battery/supercap?

[jbb]

Why not make the power supply battery powered and use the big battery to absorb the power instead of dissipating it. Or use super capacitors to store the power. ...

The downside of a battery or capacitor bank is that it has finite energy capacity.  It could be very annoying to have the load stop working in the middle of a test because the battery is full...

One thing to look out for: it can get really hard to ship lithium batteries (whether non rechargeable coin cells, LiIon, LiPo, LiFePO4...)

[danielbriggs]

I've backed your CrowdSupply thing. I can't wait, as this supply looks excellent - your attention to details is highly commendable.

However: It's currently 62% funded.
Am I right in assuming if it doesn't hit 100% funding, no one gets their PSU kits?

I really hope this isn't the case.
Is there a plan B if it fails to get there?

All the best,
Dan

[prasimix]

Yes, if it doesn't hit 100% no one gets BB3. The plan is to finally try to find a way to launch a campaign at leading crowdfunding sites such as Kickstarter and Indiegogo. It will take extra time to do this. In the case of Kickstarter, I also need an intermediary agency, since it is still not possible to start a campaign directly from Croatia.

[Ranayna]

I do not have that much personal experience with crowdfunders, but I made some observations.
I also was put off by the sharp drop of progress, but since then it grew steadily, if slowly. The initial surge were the early birds, which must have gone fast. Surprising, seeing that the discount wasn't *that* high.
Generally most campaigns see a surge near the end as well, at least if the situation is not hopeless.

Maybe we can get Dave to cover the BB3 in a Video? Does he have a H24005?

[prasimix]

I do not have that much personal experience with crowdfunders, but I made some observations.
I also was put off by the sharp drop of progress, but since then it grew steadily, if slowly. The initial surge were the early birds, which must have gone fast. Surprising, seeing that the discount wasn't *that* high.
Generally most campaigns see a surge near the end as well, at least if the situation is not hopeless.

Yes, nothing is finished yet. We've entered the most challenging phase now to make it visible to "outsiders" and I don't have a marketing team nor agency specialized in crowdfunding campaign promotions . I also didn't check how visible was Crowd Supply site two years ago (in time of H24005 funding) but I can imagine that its visibility is deteriorating (I was recently warned about that from owner of other campaign who has some "metrics" available).

Maybe we can get Dave to cover the BB3 in a Video? Does he have a H24005?

As far as I know, he doesn't. I failed to send him H24005 and that was happened again with BB3. This time I didn't manage to ensure funds for test run of modules that I can complete a few units and send them around. There was a promise from a guy who visited me this summer from Middle East willing to invest in EEZ project and get 10 units for university on which he is supposedly works but he simply disappeared, and I've lost time waiting for nothing.