different values from reg boards question

[gary h]

Hi all,

I just assembled a dc regulation circuit on a breadboard. 16V transformer supplying this circuit graciously provided by RE:



I get 400mV difference between rails. (+15.2V -14.8V)

I'm using this PS for an active crossover containing 12-14 op2134 opamps.

Is this a big deal or should I throw a variable resistor on the higher leg? If so how should this go?

Thanks,

gary

[gary h]

I just read on the ESP site that I may not post this information but I am having trouble deleting or editing this post.

[qusp]

email one of the mods and tell them you cannot expose this design, but cannot do anything about it; they will sort it out for you

[AndrewT]

ESP will gladly allow you to post a link to the diagram on his site. It's free advertising for his service.

half wave rectifier feeding a pair of IC regulators is not a nice PSU for good opamps.

[Geek]

Hi,

Easy. Use the 7915 and then a LM317 for the "+" side. You can then adjust to make for a perfect balance

Cheers!

[gary h]

Thanks for the replies,

in order:

I will contact a moderator, if Andrew is correct it'll stay. Although there is an unequivocal disclaimer at the bottom of the ESP page.

Andrew, would you kindly give a brief explanation why this is a poor choice for the opamps and what would be better, and what would be ideal. What nature of performance degradation can I expect from using this supply. I put it on my scope and the dc looks quite clean, less than 1 mV of ac. And the transformer is rated to just under 1A so I should be able to get 400mV of current for 24 134s, perhaps current is not the issue?

Gregg, that's a good idea, and I ordered an adjustable one just to see how it worked. Thanks,

thanks again,
gary

[Speedskater]

I don't think that "I get 400mV difference between rails. (+15.2V -14.8V)" will make the lest bit of difference.

[Mooly]

400mv imbalance is a non issue
Half wave vs full ? There is nothing wrong with half wave for low current supplies. It"goes against the grain" from the purist view because the ripple component is larger (because the reservoir cap is topped up only on every half cycle) and consequently the ripple fundamental frequency is at a lower frequency, 50/60 hz as opposed to 100/120 hz which make filtering more "difficult".

As you are talking of 24 ? opamps that's coming out of the realm of low current to me, and I would probably look at using a full wave bridge.

As a check connect your scope to the input of the reg and make sure that at the rated load that the regulator always sees enough to work with. Design for the mains at the lowest value too.
24 opamps, that could be quarter of an amp (or more) so heatsinking might be required on the regs too.

[AndrewT]

what are the peak currents in the supply rails?
How much of that load is met by post regulator capacitance?
How much is supplied by the regulator?
How bad is the ripple before the regulator, when supplying worst case peak load?
Will the regulator drop out on high peak loads?

half wave rectifier feeding a pair of IC regulators is not a nice PSU for good opamps.

[peranders]

Gary,

1 The regulators have +- 5% tolerance (according to datasheets). OK here

2 Opamps don't need symmetric supply voltages => you have nothing to worry about