View Full Version : peltiers on AC
71skylark
12-10-01, 06:53 AM
Due to the high amp draw of peltiers and the ineffeciency of my DC power suppply, (56A, 24VDC) which draws %75 of the amperage DC on the incoming AC. (draws 15A 120VAC to power 20A 24VDC), can I just buy a simple transformer down to 24VAC to run these peltiers? Or what is a simple way to make a EFFECIENT power supply, no matter how "dirty" or noisy the power is. peltiers don't care about noise, do they?
ButcherUK
12-10-01, 07:04 AM
Peltiers REQUIRE DC current. If your PSU is only managing 27% efficiency as you state, a new PSU might be in order. Linear supplies are almost always a lot less efficient than switching supplies. Typically a good switching supply should be able to manage 70-90% efficiency. Alternatively you can just get a big transformer, a big rectifier, and a couple of smoothing caps and have an unregulated supply. Care should be taken to make sure the voltage isn't too low or high though as it tends to vary with current draw. Typical efficiency with this might be as much as 60% depending on your transformer. Torroidal transformers are more efficient, smaller and have better EMI characteristics than iron frame types.
71skylark
12-10-01, 07:35 AM
thanks. What size and rating transformer & rectifier should I get? that was REALLY helpful info!
ButcherUK
12-10-01, 10:24 AM
Originally posted by 71skylark
thanks. What size and rating transformer & rectifier should I get? that was REALLY helpful info!
The guide for a transformer is:
Vtrans = (Vtec / 1.4) + 1.4
So for 24V you want (24 / 1.4) + 1.4 = 18.5V
Then pick the next lowest voltage.
To work out for a given transformer it's
Vout = 1.4 * (Vtrans - 1.4)
Quick explanation - The transformer puts out Vtrans volts AC. You then rectify it losing 1.4V (approx: depends on rectifier). The smoothing caps then smooth the output voltage thus going from the RMS rating of the transformer to the peak value (the 1.4* should be sqrt(2), but 1.4 works well as caps aren't 100% efficient).
As for component ratings, for a torroidal transformer you need 1.5 times the max current. So for a 20A supply at 24V you'd want a 540W transformer (18*20*1.5). For an iron frame type you should specifiy double the current.
The bridge rectifier should also be rated to double the current and double the voltage for it's PIV (peak inverse voltage). The rectifer will dissipate 1.4*I in operation. In the 20A supply this is 1.4*20 = 28W, which will probably need a heatsink. A typical rectifer without a heatsink will have a thermal resistance to air of maybe 50C/W, temps should be kept below about 100C, so for anything voer about 1W you'll need a heatsink.
Smoothing capacitors, it is recommened you use a minimum of 4000µF per amp of output current. You can get away with less but you should check the max ripple current rating of the caps and make sure you don't exceed it. If using multiple caps in parallel you add the capacitance and max ripple. For the 20A supply you'd want 80,000µF of capacitors. Lets say you can only find 10,000µF caps with a ripple current max of 8A. The minimum number you can use is 3 - ripple of 24A, the recommended number is 8. Using more is not a problem, it'll just cost you more money :) Capactiors don't dissipate energy so there is no problem with heating.
EDIT: Make sure your caps are rated for at least the output voltage, preferrably a bit more, overvolting caps will cause them to rupture - large caps will tend to explode!
For better control you can build yourself a regulator circuit. The normal design is to use a LM317 1.5A regulator and some bypass transistors. The transistors carry the current while the regulator sets the output voltage. This will give a fixed output voltage if the load varies but you'll need to dissipate quite a lot of heat from the transistors - the transistors have to dissipate any excess power to maintain the specified voltage, if your input is 30V and you want 24V at 20A out you have to dissipate enough power to lose 6V, this is 6*20 = 120W, not a small amount.
Regulation is not normally neccessary for a TEC circuit as the laod is fixed, however it does allow for variation of TEC voltage based upon temperature, you could rig a circuit so if your CPU goes idle the TEC drops to say half voltage to help prevent condensation build up. With a little planning you could make the circuit hold the TEC at a fixed tempurature.
71skylark
12-11-01, 07:33 AM
Any good online dealers of electronics parts? Should I just surround this whole rig alot of tin foil to keep EMI/RF down or is there someting smarter to do? Thanks VERY much for all the help!
Great info!
I'm building a fan controller with a small transformer to power my fans. But my problem is slightly different..
When I power up (or down for that matter), I get a loud "pop!" in my speakers. I'm using a relay to turn power on and off to the transformer, so my fans start together with my computer. I read somewhere that transformers, and specialy torroidal transformers have a very high inrush current (not sure if this was the term used).
Can this be helped or should I concentrate on "cleaning" the power to my speakers instead?
ButcherUK
12-11-01, 11:30 AM
Originally posted by Veland
Great info!
I'm building a fan controller with a small transformer to power my fans. But my problem is slightly different..
When I power up (or down for that matter), I get a loud "pop!" in my speakers. I'm using a relay to turn power on and off to the transformer, so my fans start together with my computer. I read somewhere that transformers, and specialy torroidal transformers have a very high inrush current (not sure if this was the term used).
Can this be helped or should I concentrate on "cleaning" the power to my speakers instead?
This is correct, torriods in particular have a very large inrush due to the coil being a large inductor. There's not a lot you can do about it other than try to clean the power to your speakers. A lot of PC speaker systems pop on power on as they have a built in amp with a psu and they don't clean the power.
Another question!
What happens if you draw more amps then the transformer is designed for? Is this temperature dependent?
My transformer in the fan controller I'm working on will have good airflow around it, and the biggest I can fit in a drive bay is 80VA at 15V. This is a bit low considering that I will have 5 fan outputs, each regulated with a LM317T that can handle a maximum of 1.5A each. But at present time I only have 4 fans, and none draw much power (Total of about 1.5A..)
Also, the transformer have two outputs. I understand this is normal for torrodial transformers. Should these be rectified and go throuh the caps before they're joined, or is it possible to join them before the rectifier? I vaguely remember the first aproach being the best..
Last question/request: I have abo****ely no education in electronics, but I'm totaly hooked! What good sources of information and learning are there on the net? Especially info on tranformers and caps..
Thanks!
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.