Gina

Absolutely!! Probably never know the cause as to find it would mean risking yet more RPies and I'm not doing that much as I would like to know the reason! I'll test the IRLZ44N to make sure it works on a low voltage - that's easy

Peltier TEC is 5v 20W ie. 4A and the dew heater is 70ohms so with 13.8v, a mere 0.2A. I can't see there being any problem. I've started removing components from the HAT. The new circuits are very much simpler One resistor and one MOSFET.

That looks very interesting Dave Looks like I'll have to get the Duet I would like to get back to 3D printers but I really want to get my ASC working first.

The N-channel MOSFETs I have are IRLZ44N the L indicating logic level gate control. This is the transfer characteristic from the datasheet. Seems to me that this MOSFET could be driven directly by 3.3v logic at up to 30A drain current - I'm only using a few amps.

Camera Cooler Control Circuit.

I could do this circuit. The +5v is the GPIO +5v as supplied by the RPi. Actually, the 10K resistor could be 100K or more as real protection to the RPi

The RPi uses 3.3v logic or a bit less which is not quite enough to drive even a data level MOSFET as they need 4v. I've driven those from an Arduino like you as these are capable of +5v output with pull-up. The +3.3v could be raised to +5v with a transistor and that could then drive an N-channel MOSFET. OTOH I'm sure there isn't anything wrong with my circuit with P-channel MOSFET as it has worked fine for several previous versions of my ASC (except for using a different GPIO pin - could that possibly be the cause, I wonder ). No, there must be something else wrong with the wiring on my HAT - I just can't find it. GPIO5/6 have pull-up whereas GPIO12 has pull-down and that could cause the GPIO to hover in between logic states but that would happen without the +13.8v. All the same, I wonder, it is a difference. We come back to Sherlock Holme's classic quotation...

One problem with opto-isolators is the input current required. As I recall the total current available to GPIO pins is 50mA - maybe 10mA to each opto-isolator would be alright.

That is what I'm doing and keeping the power and grounds separate with the grounds only commoned at the buck converter. All the HAT does is provide an easy way of connecting to the GPIO pins and also providing a strip board for mounting components. This is the dew heater circuit, except that I'm using a different GPIO pin. As can be seen the only connection from the +12v (+13.8v in practice) to the RPi is through the transistor. The 0v shown is the power Gnd at the input to the buck converter. The leakage current I measured as around 4 micro-amps. This has always worked fine in the past. The stepper driver modules also have separate power and data grounds and the motor output and power side is isolated from the data input and power. I guess I could use opto-isolators to avoid all possibility (almost) of power and data circuits interacting in a way they shouldn't. The camera cooling circuit is similar to the dew heater but with the power taken from the +5v side of the buck converter and the heater being the Peltier TEC cooler. ATM the +5v has not been connected to the HAT.

Here's a screenshot of the KStars FITS Viewer zoomed in to cover a bit more than half the image. Focused by hand as accurately as I could with motor removed. A few degrees rotation of the lens on its thread makes a difference to the focus. Camera Gain of zero and Exposure of 1ms (1/1000th). ASC sitting on outer window sill. Visibility is very poor today and the far hills can barely be seen.

The new RPi works with camera but without HAT.

A new Rpi 3 has arrived and I've ordered a spare - I must be stupid, I really should give up but I can't! OTOH I also can't apply "juice" to the 13.8v power wire with HAT installed or I'll blow up another RPi!

Please see my ASC Blog for all my test results. I'm stumped!

I keep peering at the HAT through my 3x magnifier and testing with my DMM but just can't find anything wrong. I connected the bench PSU to the yellow +13.8v power wire and green Gnd wires (13.8v Gnd and RPi Gnd) with HAT on duff RPi to see if there was any current - nothing significant, about 4µA even at 15v. That shows that the transistor is not shorted out as the collector resistor is 22K which would conduct around 700µA from 15v. The 4µA would be the leakage current of the transistor. So how on earth could 10v or so blow up the RPi??? HELP!! Has anyone any ideas, please?

Searching for RPi HAT I can only find them on Amazon at a fiver but delivery towards the end of January Guess I'll order one but won't solve my present problem.

I hate just giving up - makes me depressed!! And yes! I'm depressed! And I refuse to believe I'm past it! But it does take a lot longer to do things and solve problems and I guess I make more mistakes and cost myself money! With the ASC so near to working and all the structural work done on the ASC itself I can almost see the end in sight. I need to re-do the dew heater but that's a simple little job. I think I shall omit the circuitry for that for the next attempt. Oh yes there has to be a next attempt - I won't be defeated by some silly electronics!!

I think my ASC project is doomed

Looks like it's time for a different project

13.8v Gnd was fine and I went on to connect the +13.8v wire. That wasn't I now have yet another cooked Raspberry Pi So what do I do now??!! This is getting (got) expensive!! The only thing I can think of is to buy another HAT and wire that up. The only difference from when this HAT was working fine was that I added a second position for an A4988 stepper driver module, removing some components and moving the first module position to make room for the second. The dew heater and camera cooling circuits have not been touched. I just don't understand it - there is no circuit to the RPi except via high value resistors and the drive transistors which are in turn controlled from GPIO pins and +3.3v. The resistors and transistors are in the dew heater circuit.

Next to shut down again and connect the 13.8v Gnd.

Passed that test

Can't put it off any longer - today I plug the RPi in to the HAT and power up and fingers crossed... Just the HAT to start with without external power and no motor driver modules. One small step at a time.

I have an Ethernet shield for an Arduino Uno but it can't go on the present Uno as it already has the data logger shield on it and neither are "pass-through". But I could add another Uno to the I2C network and plug the Ethernet shield into that. Or maybe connect the Uno to an RPi if I work out how that might work I have a CAT6 cable into the observatory. Guess I could put my gigabit switch in the observatory as I don't think I need it indoors.

Now have RPi attached to its bracket on the ASC and wired to the buck converter which is being supplied with 13.8v via its long power cable. Everything still working. No HAT on yet. I think that's as far as I'm going tonight - going to stop while ahead