[SonOfShrink]

I have built my STEPS as per the Tangent web site.
Australia's AC Voltage is 240V.

I had already ordered the 25VA 2X 15V transformer as per Tangent's instructions, by the time Guzzler notified me that the regulator may run a bit hot. I wish to use the STEPS to power an M³ with AD8610 op-amps and thus need the output voltage to be 24V.

The LM317 voltage regulator runs very hot. I am concerned about it overheating the other components. I have insulated it (and used silicone grease) from the heat sink.

My voltages are as follows:

NO LOAD:
test points 1 and 2 - 38V
test points 3 and 5 - 50V
test points 4 and 5 - 24V

M³ connected and powered up:
test points 1 and 2 - 34.7V
test points 3 and 5 - 40.8V
test points 4 and 5 - 24V

Which transfornmer should I use, the 25VA 2X 12V one ?

Is there another way to make the regulator run cooler ??? I realise that it is dropping way too large a voltage. The M³ draws about 0.3A

Should I change my op-amps to OPA627Ap and OPA637AP and run my power supply at 36V? This would make the regulator run cooler ???

 

[steinchen]

Quote:

Originally Posted by SonOfShrink Which transfornmer should I use, the 25VA 2X 12V one ?

for 24V output: yes

Quote:

Originally Posted by SonOfShrink Is there another way to make the regulator run cooler ???

dial in a higher output voltage, but you have to use other opamps in this case

Quote:

Originally Posted by SonOfShrink Should I change my op-amps to OPA627Ap and OPA637AP and run my power supply at 36V? This would make the regulator run cooler ???

36V is way too much and I doubt your psu is able to deliver it. Nevertheless keeping the 2x15V transformer, adjusting the trimpot for 30V output (you probably have to change a resistor) and using opa627/637 is a reasonable solution

 

[n_maher]

Or just do what I do and use a bigger heatsink. Digikey carries others that will work fine and keep temps much lower. Also, you should be able to dial the steps up to 26V and still only get 24.X at the opamps which is fine for the 8610s.

Nate

 

[steinchen]

neat and useful tool: Power Supply Parameter Estimator
http://www.tangentsoft.net/elec/ps-est.html

@tangent
you may consider adding a link from the steps part list and part selection guide to the PSPE

 

[Teerawit]

And another thing you COULD do is lower the biases on the M3.

 

[dsavitsk]

Or you could put a voltage dropping resistor in series with your amp takiing some of the strain off of the regulator. So, set the STEPS at 30V. You then need to drop 6V. Since the amp draws 0.3A

V = IR

6 = (0.3)R

You need a 20 ohm resistor. Further, the wattage rating of the resistor is the resistance times the current squared. So, you need a greater than 2 watt resistor to drop that much voltage. I would use a 5 watter (or at the very least a 3 watt) to be on the safe side.

As a last option, get a little computer fan -- preferably one that draws a very small current. Use the same formula to drop its voltage to someting like 5V so it will run silently and point it at the regulator. You do not need very much air power to keep the heatsink very cool. I do this to a stack of DAC chips that were running so hot I couldn't touch them. With a 5V fan on them, they do not even feel warm. But note, the more current the fan draws, the more heat it needs to power away from the heatsink, so do be sure it is a low current fan (I have seen them vary from 0.012A to 0.5A for the same sized fan.)

 

[Teerawit]

SonOfShrink, have you measured the temperature of the regulator/heatsink when connected to the M^3? Do you get around 55 degrees Celsius?

 

[amb]

Quote:

Originally Posted by dsavitsk Or you could put a voltage dropping resistor in series with your amp takiing some of the strain off of the regulator.

I would advice against that, because that would increase the effective output impedance of the power supply as seen by the amp, and adversely affect dynamic voltage regulation. If one must add a series resistor, the right place to do it is between the power transformer and the input of the regulator, but this is not possible with the STEPS without hacking up the board.

 

[Teerawit]

At your current power supply setting, your regulator should be around 60 deg Celsius under load. Since the LM317 can operate up to 125 deg C, it won't be adversely affected...but....

I am curious - is it safe to keep it at this temperature? How long can the LM317 survive at 60 deg Celsius?

 

[dsavitsk]

Quote:

Originally Posted by amb I would advice against that, because that would increase the effective output impedance of the power supply as seen by the amp, and adversely affect dynamic voltage regulation. If one must add a series resistor, the right place to do it is between the power transformer and the input of the regulator, but this is not possible with the STEPS without hacking up the board.

How about adding another cap after the dropping resistor to make it an RC filter?

 

[amb]

Quote:

Originally Posted by dsavitsk How about adding another cap after the dropping resistor to make it an RC filter?

On the amp board there is a bank of rail caps so such a filter already exists. The caps do reduce the impedance, but it's best to not raise it in the first place, especially at the output of the regulator.

 

[SonOfShrink]

I had a look at very small fans and they cost heaps in Australia from RS Components. So, I will substitute a 12V + 12V transformer. I will keep the other one for my next project. Thansk for all the help.

 

[alphaman]

Yeah .... I ordered the STEPS pcb from tangentsoft back in 2009, built it, tested it with a DMM and oscope, wanted it to pump 24v into my PPVv2. But the damn regulator -- 317 or 1086 -- got very hot, even with a large heat-sink. What a joke .... and (rightfully) the dodgy stock design has rec'd much criticism on this and other forums. Too bad tangent spent overtime on the presentation and making the now-Archived web page look pretty (and not enough time on the topology).
Proof that I built a STEPS .... I guess it proves I was a nooooooooob, young-n-dumb, or just plan naive ...

I think folks who think that the same designer who was ONE contributor to Pimeta and PPA, has somehow pulled another winner outta his hat is this reg. is in for a BIG surprise.
Damn ... what a waste of time and $$. I guess that 'splains why Tangent abandoned the project to the bowels of Archive.org. Too many 'planits comin' in, methinks. Uggghhhh!
https://web.archive.org/web/20100611215018/http://tangentsoft.net/elec/teps/