Anushri (and Shruthi) 7.5v power supplies - the good ones are bad, the bad ones are good

I recently bought a somewhat more expensive, good 7.5v wallmart for my Anushri. This one was not an ‘el cheapo’, but ‘made in Germany’ and ‘allowed for medical applications’. The idea of using a 7.5v power supply is that the regulators don’t get too hot, even though that does not have to be a problem (heatsinks etc.).

I already had some cheap 7.5v power supplies, which worked well, but this fancy one did not ! That is, it took about a minute or so before the Anushri started to become usable …

Out comes the multimeter, and the power supply turns out to be well made: it puts out 7.6 to 7.7v (loaded).
Measuring the ‘el cheapo’ 7.5v ones (with the Anushri connected, switched on, and working well) revealed (as has been said earlier in this forum) that these put out quite a bit more than promised: 10v !

But these 10v produce a nice +5v and -5v on the two rails inside the Anushri, whereas the 7.6-7.7v results in a +5v on the positive rail, and -4.2v on the negative one.

Hmmm. A little adding up of the voltage drops from the power connector to the rails is the answer, I think.

The 1N400x protection diode has a drop of around 0.7v, and the 7805/7905 regulators around 1.5v to 2v.
That means that the +5v for the positive rail is (barely) reached after these two components.

But for the negative rail the LT1054 has to be taken into account as well, which not only switches polarity, but looses almost a volt as well.
For my 7.7v (actual) power supply, the protection diode takes away 0.7, so the LT1054 gets fed with +7v, and produces about -6v (measured). The 7905 turns this into -4.2v …

So I guess one should not get a 7.5v power supply that actually turns out anything below 8.5v ! Most of the cheap 7.5v don’t do this anyway, so are fine, but for ‘good’ power supplies one should stay above 9v for a good negative rail voltage.

Keep in mind that the excess energy is dissipated into heat……

Of course, that was the whole point of going for a 7.5v wallmart anyway - but as shown, if the wallmart actually produces anything near 7.5v, you end up with a (roughly) -4.2v negative rail inside the synth.

The regulators don’t get too hot with the cheap 7.5v (but really 10v) wallmarts. I put a little heatsink on the 7805 anyway.

Maybe you better swap your 1054 for one with better specs - problem might be that the voltage drops to much so the 7905 has not sufficient overhead.

Ah … there are different ones ?
This one is a LT1054CN8 …

The datasheet I found mentions a voltage drop of about a volt, though (it lists a 1.1v drop for a 100mA current).

Googling a bit more, the CN8 part seem to only relate to the packaging.

This is interesting to me since I have something similar going on with the LT1054 outputting a -9.5V instead of the expected -12V when fed with a 12V supply. I checked all the cap orientations & solder joints for the PSU and everything checks out.

I tacked an extra LT1054 onto an order I was already doing, so I’m going to try replacing it and see what happens. I’d be interested to know if replacing yours fixes your problem (assuming you’re going to try it).

Is the LT1054 outputting -9.5v, or is that what is left after the 7905 regulation ?

The protection diode takes away 0.7v from the 12v at the PSU, so the LT1054 will be fed with 11.3v (perhaps you can measure this ?). Losing 1.8v (from 11.3v to -9.5v) seems a bit much, though.

PS: in my case it is not really a problem, as the LT1054 works as it should (the 1v voltage drop is according to spec). I just can’t use the ‘nice’ 7.5v instead of the ‘cheap’ 7.5v ones (which actually output 10v). Not a big problem, but good to know.

That’s what the LT1054 is outputting, measured on pin 5. I don’t specifically remember what the 7905 is outputting, I’ll have to look at that when I get home to see if I’m also seeing less (or more) than -5V.

I agree about losing 1.8V, at first I was suspicious of the orientation of my caps, but everything checked out. I’ll try replacing the LT1054 and see what happens.

Thanks for reporting this.

I guess in future communications, I’ll restrict the choice of input voltages to “9V”. 12V and there’s the risk of exceeding 15V during transient events; 7.5V and indeed there could be too much of a voltage drop to allow the 7905 to correctly operate.

I also like the idea of using a 7.5V PSU because of less energy turned into heat.

I had exactly the same issue with the same measurements. I first used a L7905CV: It had 6.22V in and 4.30V out. Next I replaced it with a MC7905ACTG from mouser and voila: 6.22V in and 5.00V out exactly. Larsen explained that the new 7905 is designed to work with a lower-than-usual voltage drop (an LDO regulator).

Interesting ! I guess you meant to quote negative voltages, though, as the positive ones are not the problem.

So going with the ‘next generation’ 7905 helps … that is an option. But just about getting to -5V exactly might be a bit uncomfortable ?

But I guess going for a good 9v wallmart or a ‘cheap’ 7.5v wallmart (of the type that gives over 9v anyway) is the simplest solution.

Looking up the MC7905ACTG part at Mouser, the minimum input voltage is still quoted at -7V …
(of course max and min should be inverted for these negative voltages, but hey).

Yes, I meant -5.0V of course. Otherwise my YM Shruti might have other problems if there were +5.0V coming from the 7905.

I don´t know exactly why - but I have both voltages in range now as intended. I use an adjustable 600mA switching power supply at 7.5V.

Perhaps Mouser wants to play safe with those quoted specs …

But I think Olivier is right to restrict the recommended value to 9v, as for 7.5v it depends on the actual wallmart (what it outputs) and/or your choice of components (loopino’s nextgen 7905 seems to (barely) do the trick, my older one does not).

Loopino, what does your wallmart actually output ? Usually not the same as what is stated …

Edit: clicking on the datasheet link at Mouser shows that for loopino’s part the voltage drop is 1.3v.

I´ll try to have a look at it this evening.

For the future builders it would be safe to prior test the psu with an artificial load - not the Anushri itself. In the manual can be written how to do that. It is possible to establish a test like - take a 12V/5W G4 or sofita bulb, connect it to the psu. When it lights, measure the voltage across it. The measured voltage must be in the 8,5V-11V range while the bulb lights.
Or something like this.

The poor beginner guys that have no G4 / sofita at hand………