Jake_G wrote:II'm with you for the most part, but what is the 7805?

Not sure if you were asking what a 7805 is or what Assaf meant by "holding on for dear life." Or both.

Anyway, it's a 5 VDC kinear voltage regulator. Typical part number is LM7805, but there are may variations that carry the 7805 designation (7905 is the negative voltage version.)

Feed it, say, 13.8 VDC and you get a well-regulated 5 VDC out. The circuit should have an electrolytic or tantalum cap on the input, somewhere around .33uf, if it's far away from the power supply (though every circuit I've seen has it one) and a .1uf output cap to improve stability and suppress transients.

Assaf is saying that 6.7 VDC is quite low input voltage for a 7805. I'll say. The data sheet says the input voltage should be 2V above the output voltage. Holding on for dear life, indeed.

Though a lot of modern 7805 like voltage regulators can live with a lot less than 2V.

Ira

Yes. LDO (low dropout) regulators do exist. And yet the energy wasteful 78XX series continues to be popular in areas where an energy star isn't a design goal.

Obviously switching power supplies are better energy-wise but don't have the longevity of linear ones (unless designed for that - industrial ones are a pretty penny and go to extraordinary lengths to sharpen the FET pulses - for efficiency sake - but to flatten the pulses by the time They are pulled from or charged at the reservoir caps - for longevity Sake- - read as expensive caps, lots of them and inductor galore).

The energy saving, though, is substantial. In the 3d5 the energy wasted on the 7805 is more than 50% (assuming 12V in 5V out). A nice touch is the 12V relays powered by open collector outputs. So that the current comes directly from the transformer power and doesn't flow through the 7805.

Thanks all. I appreciate the schooling. Similar to my experience with the boiler fill circuit, a big goal of mine here is to understand how the 3d5 was designed to work. In this case, fixing the problem is actually priority #2, since the machine works just fine right now. In fact, it hasn't reset since since the morning after swapping the solenoid coils. I really enjoy the bn learning process here and as a side benefit, we should be able to document a root cause and Mike can get his machine fixed, too.

You still with us, Mike?

Of note on the 7805(s), is that there appears to be 2 of them, in series. Here is a decent view of the rectifier and regulators:

You can see the two AC traces feeding the rectifier from the output pins of the transformer and the positive trace that feeds unregulated DC to our regulators coming out of the lower left pin on the rectifier.
C1 is out of view following the fat trace up and to the right and is 1000uF.
The terminals in the top left corner are for the coil of the pump relay and the top terminal is fed directly from the rectifier; the rail feeding the other relays is on the back side extending from that pin.
The long run downwards feeds the 4 pin socket labeled "ATE". Left to right on that socket are the unregulated DC, the negative watchdog output (3Y from the IC), the positive watchdog output (1Y) and finally a ground pin. What is this socket for?
Finally, we feed the first regulator just below the rectifier with a drop leg off of the run that feeds C1 (and also the ssr for the coffee boiler).

You can see here that the output of the first regulator (left pin of the upper 3 pin group under the rectifier) feeds the input of the second regulator. There is a ceramic cap between this intermediate voltage and the center ground pin of the second regulator. There is also a ceramic cap on the outfeed of the second regulator tied back to the same center ground pin.

The right pin of the lower regulator is the source of Vcc for the board and immediately feeds C2, which is 100uF and grounded through 3Y on the watchdog. Presumably, if you actually lose voltage, pulling C2 keeps it from discharging through Vcc and ensures that the logic of the board is truly down for the count until voltage is restored and the tantalum caps feeding the watchdog circuits are both recharged.

Of slight interest in the upper right hand corner are the carefully regulated current sources for the twin PT1000 RTD circuits that feed the PIDs for temperature control.

I write all of this to say "why two regs?". Is the first stage not a regulator? They both have the same form factor and the same bulky heat sinks that obscure any identifying markings. It doesn't appear that the intermediate trace between the stages is used anywhere on the board, and dropping the voltage to 5V and then sending into another regulator seems like a decidedly bad thing to do...

Thoughts?

Cheers!

- Jake

I'm not sure the second regulator is a 7805 -- I can't see the label in any of my photos of the board. It could be a 7812 for the 12 VDC supply, though the form factor doesn't quite look like a 7812, the form factor of which is usually (though not always) the same as the 7805. It looks wider and thinner in my photos. Could be a higher amperage version.

I know there's 12 VDC available on the board for the relays and probably the button lights, and it may feed the 5 VDC regulator. I used the 12 VDC to power a steam button mod (small relay to switch AC to a solenoid and LED in the switch.) The relays and lights don't necessarily require regulated voltage, but they do get switched through ICs and I don't usually see unregulated voltage switched that way.

Assaf, are you sure the relays work directly off the transformer output? I'd think they use regulated voltage.

Phew... I'm the one who originally posted the thread, then it kind of took off and went way over my head pretty quickly. Things with mine have been acting OK but something strange is going on with my steam boiler and keeping pressure at the proper bar. So far, I haven't seen the machine reboot since I've posted it.

I will say though, when my machine did reboot, it would automatically start up again. The video that was posted showed you had to turn it back on manually. While not positive, I think I'm on an earlier firmware version too (maybe 1.13?).

Sorry I don't have much more to add... While I can solder and replace components, I'm not great at knowing what electrical components do or how they work. Nonetheless, I'm glad that the start of this thread may be able to help somebody out with this issue, and maybe help me in the future when my machine starts doing it again.

Oh, and secondly... Jake, AssafL, and anybody else. These threads are of huge help, between this one and the one Jake did on rebuilding his machine. When the time comes, and I think the time is coming soon, to rebuild my machine, these will help out more than you know.

Have you ever thought of combining all this knowledge into an unofficial troubleshooting book people can buy? It would be great if there was a one stop shop for troubleshooting and working on the GS3. I don't know what the latest serial numbers are, but I think there's many people out there that own one and don't mind rolling their sleeves up, but just don't know how. There was a Linea 'orange book', but all the technical information on the gs3 has come from forums like this. I don't know, just a thought.

wiz561 wrote:Things with mine have been acting OK but something strange is going on with my steam boiler and keeping pressure at the proper bar.

This is likely a serious problem and needs to be addressed. You PMed me about it and I gave you some suggestions, but I think it would be helpful to both you and other GS/3 owners if you post the symptoms here on HB under a different thread.

FWIW, there are a number of things that could cause the problem, including a bad temperature probe, boiler overfill, leaky autofill solenoid, leaky vacuum breaker, etc.

Well, curiouser and curiouser... IC1 (the closer one to the transformer) is a 7812: A 12V regulator. While IC2 is a 7805.

Dick - the relays are not powered by the regulated 12V but by the 13.6V of the xformer. You can easily trace the route from the rectifier to the large motor relay. From there it flips to the ground side and works its way to all relays.

BTW - The 7812 is hanging on for dear life - but not successfully. I measured 11.7V.

I am stumped...

Even if they needed the 12V (lets say for the RS232 - but even that is TTL 5V level on the board) - why route the 5V through it? Even if the 12V regulator takes 2V off the 5V regulator - it removes a measly 3W (2V x 1.5A). It isn't worth the double losses and double large heatsinks...

A hunch/something to check. Perhaps the 12V is somehow connected to the watchdog? serving as a crowbar to throw the system if the DC voltage goes too low?

"ATE" (I think) stands for Automated Test Equipment. It refers to the test bed used to test the boards. There was a case study on the NI website. See Interesting case study from Gicar Srl. It seems to have been taken down.

Given the pinout of the ATE connector it seems to me that it is the way power is supplied and resets tested before the automated "bed of nails" starts running the tested features.