EEVblog #291 – USB Power Supply Part 2 – The Housing

[FenderBender]

As you say, a wobbly back to an enclosure is a total non-starter when using tactile switches. Maybe this would be a good excuse for Dave to investigate the ins and outs of injection moulding manufacture? Even if it is just rubber "bumpers" that he can fit over each end of the enclosure (although the asymmetric lip on each half is kinda clunky too).

Rubber bumpers would be much more professional than stick on rubber feet, I agree.

[NiHaoMike]

The link (to the video) on the front page is wrong.

[EEVblog]

I understand the concern with the housing. However I have yet to find another suitable case that meets the credit card size form factor with removable end and at a comparable price.
If you know of a specific one that does, please point it out.
I expect the PSU to be handheld a lot, and possibly operated while in the hand, rendering the curved bottom moot in those cases.
It's a compromise I thought was worthwhile given case availability and price.
As usual, people like to throw around all sorts of solutions, but they haven't really looked in to the finer details and costs of it, and that's when you often "come a gutsa".

I am actually considering the PCB top like the uCurrent, with components underneath and touch switches, negating the extra cost for the custom plastic. But the devil is in the detail. For example, that instantly doubles the price of the micro. But that might come out in the wash.

Dave.

[mikeselectricstuff]

I am actually considering the PCB top like the uCurrent, with components underneath and touch switches, negating the extra cost for the custom plastic. But the devil is in the detail. For example, that instantly doubles the price of the micro. But that might come out in the wash.
Dave.

Is that for touch peripherals or just extra pins? For simple PCB touch buttons which have just resist/silk over them you don't really need touch peripherals as you can easily do it with just pins, measuring discharge time over a few uS with ~100K pullup.

[EEVblog]

Is that for touch peripherals or just extra pins? For simple PCB touch buttons which have just resist/silk over them you don't really need touch peripherals as you can easily do it with just pins, measuring discharge time over a few uS with ~100K pullup.

I've speced in an ATtiny48AU which is $0.86 in 1Kqty. There seems to be a significant jump up in price when going to even moderately more feature rich micros.
Could probably do as you say, but I've never tried it that way. Would require some testing I suspect.

Dave.

[pickle9000]

Personally I'd like to see the plastic much stronger and safer. Yes it may get a scratch or two but it is a tool. If your worried about it get a sock to carry it. It's not a real tool without some burns and scratches on it.

If you have the space for a couple more buttons I'd love to see a couple of one touch presets and perhaps a simple keyboard lock. The actual size looks very good and while knobs are nice on a bench unit, they could cause damage on a small rig like this.

I'm really curious to see how the display was going to be set up.

...mike

[EEVblog]

Personally I'd like to see the plastic much stronger and safer. Yes it may get a scratch or two but it is a tool. If your worried about it get a sock to carry it. It's not a real tool without some burns and scratches on it.

Yes, scratches are inevitable I guess. And if it was a PCB front panel, it would just get scratched as well I suspect. Would definitely need the tougher photo-imageable component overlay like CircuitLabs (NZ) do in this case.

If you have the space for a couple more buttons I'd love to see a couple of one touch presets and perhaps a simple keyboard lock.

No room, sorry.

I'm really curious to see how the display was going to be set up.

You'll see!
I'm still not sure how good it's going to be, but it's very cheap.

Dave.

[markus_b]

What about the same box as the uCurrent ?

You could mount the output poles on the sidewall. All the other components go on the PCB.

Yes, the resulting device is thicker than you would like, but it will have the 'EEVblog-look' (tm). Also, when I carry a power supply then I'll carry allkinds of other junk as well, so I'll take a big laptop bag.

[G7PSK]

How about polycarbonate instead of perspex it can be thinner and is scratch resistant.

[EEVblog]

How about polycarbonate instead of perspex it can be thinner and is scratch resistant.

I'm pricing both out now.

Dave.

[robrenz]

How about polycarbonate instead of perspex it can be thinner and is scratch resistant.

I'm pricing both out now.

Dave.

Both are available with SAR  (super abrasion resistant) coating that is not much more expensive.

[TopGunPk]

Hi Dave.

Good housing design, specially when you can use 1 case to make 2  (Top and bottom cover).

One thing though - since this is a USB powered PSU, do not you think the 7-segment LED displays will chew up a lot of current from the USB bus ? Considering minimum and maximum current draw, assuming you are not doing any fancy multiplexing on the display segments.

If the display shows "1", it means only 2x LEDs are lit up, while worst case scenario could be all displays showing "8888". That would not leave much room for the DC-DC converter. Or am i missing something here ?

Regards,
Zeeshan

[markus_b]

Was browsing for boxes, here an alternatives:

Hammond 1551 Series (around $2 a piece): http://www.hammondmfg.com/dwg9.htm

[pickle9000]

I'm really curious to see how the display was going to be set up.

You'll see!
I'm still not sure how good it's going to be, but it's very cheap.

Dave.

I'm thinking you will cut slots in the pcb for the segments and span them with an LED. Not sure if that is cheap compared to an actual 7 seg display but would definitely be your style. I think it could look pretty nice and start a whole new way if using pcbs. Before you know it you will be arranging the components on the board so that they resemble Sagan. That would be cool.

...mike

[Zad]

Some of the Hammond 1593 series have nearly symmetrical halves (again, one half has countersunk holes) but, crucially, the halves are  flat.

http://www.hammondmfg.com/1593grp.jpg

http://www.hammondmfg.com/1593ptbl.htm

Some have battery doors too. It might be possible to design it so that a talented hacker could swap enclosures and install 4xAA cells,  and run it off those.

[LaurenceW]

I think Dave's video raises a couple of really important points, for anyone who is considering making a low-volume product (maybe a few hundred or less). Any custom case (especially any custom mouldings) could easily end up costing you more than the elctronics! You want something cheap, and yet it doesn't want it looking like something a seven-year-old threw together, and that is this.

Start with the case first, then figure out how to get your electronics in to it!

The second point, which he only briefly alludes to, is designing your front of back panel at the same time as your PCB. Choose a sensible colour (like a black solder mask with white screen printing and viola - a professional looking part, for pennies.

I used exactly these two tricks for a bunch of miniature nixie clocks that I have designed. The result (I think), is quite pleasing and professional.

[8086]

I think Dave's video raises a couple of really important points, for anyone who is considering making a low-volume product (maybe a few hundred or less). Any custom case (especially any custom mouldings) could easily end up costing you more than the elctronics! You want something cheap, and yet it doesn't want it looking like something a seven-year-old threw together, and that is this.

Start with the case first, then figure out how to get your electronics in to it!

The second point, which he only briefly alludes to, is designing your front of back panel at the same time as your PCB. Choose a sensible colour (like a black solder mask with white screen printing and viola - a professional looking part, for pennies.

I used exactly these two tricks for a bunch of miniature nixie clocks that I have designed. The result (I think), is quite pleasing and professional.

That is really nice! Where might one find that case? Looks snazzy 

[LaurenceW]

8086, these cases, made by Serpac

http://uk.mouser.com/Search/ProductDetail.aspx?R=A-31-BLACKvirtualkey63500000virtualkey635-A-31-B

I actually bought mine from Mouser, but they may be available directly in the UK, now (although the "direct from UK distributor" prices are STUPID! Guess they DON'T want the business... http://www.serpac.co.uk/enclosures/a-series.htm[/url])

You can also get moulded clear plastic windows, though optically these are not as good as a piece of acrylic sheet. But then you've got to factor in the cost of someone cutting the acrylic sheet to a non-rectangular profile, of course.

[Jope]

One (or the) manufacturer of this cases is the Taiwanese company Gainta Industries: Gainta website
The case Dave wants to use is one of the 400 series.

[typeglob]

I wonder if Dave's going to replace the 7-segment displays with SMD LEDs mounted in a similar pattern. I assume its going to be (partly) machine assembled anyway.

Doing away with the housing for the LED displays might require lowering the brightness of the LEDs (so as not to blind people, as you'll be losing the diffusor) and maybe mounting two LEDs per segment.

[Short Circuit]

When a countersunk screw is installed in the bottom hole, you get a flat surface.
Decent quality bumperstops actually adhere very well to such surface and won't come off.
Also, the screw is sightly recessed, which only improves durability of the feet since they have nowehere to go horizontally.
Simple solution for the top half is to duplicate the holes with a countersink drill on bench drill with adjustable stop.

[Zad]

Using the LCD from the battery powered PSU project would save on inventory and improve the BOM scalability.

[hlavac]

I did a little more precise estimation on what could be possible with the USB power supply as a little competition to Dave's design
It goes the comlpex and expensive route This is what I got so far:

My imaginary design has two isolated voltage outputs, one fixed 5V and one variable 0-20V with total combined output power of roughly 1.3W, and an isolated logic level RS-232. I may even provide some GPIO to allow for creation of i.e. a simple GAL programmer.
Both DC-DC converters are powered from the USB side to avoid double conversion loss.

I have put ATmega32U4 on the USB side to handle 2 3digit 7segment LED displays (found nice low current ones which take only 2mA@2.1V per segment) and some controls (not decided yet, buttons for now, probably will attempt two capacitive sliders but I/O is getting scarce on the USB side).
It does have capability to do a proper USB handshake and thus negotiate power consumption, has USB remote control interface for the supply and provides virtual serial port that goes to the isolated side as isolated 5V TTL level RS-252.

There is an isolated serial bus between the two microcontrollers that 1) controls the PWM that sets voltage and current levels on the isolated side, 2) provides for data path of the measured voltage/curent back to the ATmega32U4 for display on the LED displays and communication over USB to the optional control panel application, 3) allows configuration of the baud rate on the isolated serial bus to the device and 4) passes serial data from the virtual serial port to second UART of the ATmega164A on the isolated side.
ATmega164A on the isolated side is doing current and voltage measurements with the internal ADC and sends data back and forth over the middle bus (which runs faster than the virtual serial to allow for internal communication in addition to the user data).
I have two 78S40 on the USB side to do the power conversion, voltage is measured on the isolated side with comparators and result is passed back thru two optocouplers (on=too high). Variable voltage is compared against PWM generated threshold voltage generated by the isolated side MCU. This part is too simple and imprecise for now, need to find a way to generate PWM based on a reference not just output of the MCU which depends on its supply fluctuations. Maybe I will just power whole MCU from a voltage reference of some sort.
What I'm currently missing is a current limiter on the 5V output. It will be necessary so I dont starve out the isolated side MCU by excessive current.
Also I need to figure out how to combine the current limitation of the two converters together into a common power budget without limiting them both to a fixed current value. I'll probably just stick a current limiter between the USB power and the rest of the circuit for some loss in available power.
I'm also thinking of a 5V wall adapter connector to allow standalone operation without USB and with more power.

So, what do you think, overcomplicated?

[Bored@Work]

So, what do you think, overcomplicated?

The 78S40 really hurts. It is a 40 year old switching regulator, basically a MC34063 with an additional opamp. If you plan to go the expensive route, then use modern switching regulators, not that old stuff. It is hard to make an adjustable SMPS, especially, if the input voltage (5V) is within the output voltage range (0 - 20V), and an 78S40/MC34063 makes it even harder.

To get an idea what it takes to make an adjustable SMPS, see mike's recent teardown video. And pay attention to the part where he talks about the weird behavior of that SMPS under special loads. That's the SMPS oscillating, because the control loop is not properly compensated for that particular load. And that happens with an SMPS specifically designed to be adjustable. It is not just some bare minimum circuit taken from an app note with the feedback resistors replaced by some PWM interface.

As for the rest, two micro controllers communicating over an isolated serial link is a standard technique that was even used before there were micro controllers, but only micro processors. I would not use TTL-level RS232 for it, but SPI (or I2C). The USB side being the master, and the isolated MCU being the slave. And I would only have one isolated bus between the MCUs. The isolated slave doing the actual PWM, while the master just communicates the desired values to the isolated slave over the bus when there is a change. And the master regularly polls the isolated slave regarding the current values (voltage, current), to display it.

Once you have the isolated SPI bus (or I2C), you can add more slaves to the isolated side of the bus. E.g. instead of using PWM, you could add some DACs to the bus on the isolated side for voltage and current adjustments instead.

I would use a supply voltage of 3.3 V or 3.6 V on the isolated side, not 5V. Generating a stable 5V from the USB voltage is more effort, because the 5V aren't guaranteed. The USB standard says something like 4.5V to 5.5V (you would have to check, I am currently to lazy to look it up in the standard). Using 3.x V means you can just use a modern buck converter.

And because of that supply voltage I would drop the fixed 5V output. Either make that 3.x V, or skip the feature.

[hlavac]

The 78S40 really hurts. It is a 40 year old switching regulator, basically a MC34063 with an additional opamp.

Yeah, but It's all I can get here for cheap. I really need to try Farnell or something. 40 years behind is literally where local component stores are here in Czech Republic
I liked it because it was kind of a set of stuff you need for SMPS, not a magic black box. And unlike the MC34063 the voltage reference is not hardwired to the comparator but connected to a pin.

What is wrong with it, except for being archaic? Bad switches? Low efficiency? I would be pretty contempt with about 80%.

I would not use TTL-level RS232 for it, but SPI (or I2C). The USB side being the master, and the isolated MCU being the slave. And I would only have one isolated bus between the MCUs.

Well the point of RS-232 was it only needs one line/optocoupler per direction. Optocouplers are amongst the more power hungry components in the design, and I'm trying to shave milliwatts here.
I got traffic in both ways as the RS-232 port is bidirectional and I can set the voltage in one direction and read the measurement in other.  Rest is really as you describe.

Once you have the isolated SPI bus (or I2C), you can add more slaves to the isolated side of the bus.

I could just hook that up to the isolated MCU no?

I would use a supply voltage of 3.3 V or 3.6 V on the isolated side, not 5V. Generating a stable 5V from the USB voltage is more effort, because the 5V aren't guaranteed. The USB standard says something like 4.5V to 5.5V (you would have to check, I am currently to lazy to look it up in the standard). Using 3.x V means you can just use a modern buck converter.

As the topology will be flyback thru a 1:1 transformer not buck because of the isolation, that should not be a problem?
Or how can you do isolated buck? I don't think you can.

I wanted 5V so that I can use it on the external RS-232. Most of my designs are still 5V. Maybe even switch the isolated MCU power rail between 5V and 3.3V depending on what I need for the RS-232.

And because of that supply voltage I would drop the fixed 5V output. Either make that 3.x V, or skip the feature.

I'll think about it. It does seem to complicate things (like I can not use low side current shunt on GND for example as it is shared with the 5V). I'll see if I can figure out the high side current measurement, if not I'll have to drop the 5V.
But I'd really like to keep the isolated RS-232 to the powered device...