| JoeBob |
| Well, the highest rated bridge I can get is 35A and the highest load possible in my amp is 30A, now that's a little too close for comfort, so my questions is this... I'm using two 750VA transformers, two 61000 uF caps and it's a five channel amp, everything else is done except for this. Instead of running the two secondaries of the transformers in parallel and then to the rectifier then the caps could I run each transformer to it's own rectifier and then the outputs of the two rectifiers in parallel then to the caps? Or any other way to use two rectifiers so that I'm not so close to it's rated limit? Thanks alot for any replies. |
|
|
| audionut |
Each transformer could have its own bridge rectifier. One transformer and bridge rectifier for the plus rail and one transformer and bridge rectifier for the negative rail. Each bridge would have less of a load.
You can go to the tnt audio sight and find a good powersupply article which talks about the merits of having separate windings/transformers and separate bridge rectifiers for each rail...also one of the recent Pass Labs threads also talks of the advantages of having two windings and two bridge rectifiers.
good luck! |
|
|
| Joe Berry |
joebob,
Bridge rectifiers usually have both average and peak current ratings, where the peak rating may be around 10 times the average. So, a bridge rectifier with a 35A average current rating will probably handle 350A or so under peak conditions. If the parts catalog doesn't include both ratings, you can get a databook (or maybe download a datasheet) with full specs for the part in quesiton. |
|
|
| JoeBob |
| Oh, I see. I'll go check for a datasheet. I'd be alot happier if I knew it wouldn't blow up. |
|
|
| blmn |
JoeBob,
When using these diodes near of their capacity you must be aware of their heatsink needs (even far from this capacity too). Please, consider the average current over them and the voltage drop (you have this information in the datasheet).
regards |
|
|
| paulb |
At such high currents, another option is to use individual rectifiers instead of a bridge. Be very aware of the heatsink requirements: if you're pulling 10 amps continuously, a bridge might have to dissipate something in the neighborhood of 15 or 20 watts.
I prefer the option of splitting it up so that one transformer is for, say, the positive and one for the negative rail, or perhaps split them up by channels. |
|
|
| JoeBob |
| So I'd hook one transformer to the bridge and only connect the negative from that bridge, and the other transformer to another bridge with only the positive connected? Sine my transformers have two secondaries each I'd hook them up in parallel to the bridges? |
|
|
| subwo1 |
| Or, you may be interested in hooking each secondary up to its own bridge, like you were saying at first. Then hook 2 negatives and 2 positives to ground like you said in your last post. |
|
|
| Jamie F |
JoeBob
I'm using a single 35A rectifier for 10 LM3886's and it generates no appreciable heat.
Jamie |
|
|
| paulb |
| quote: | Originally posted by Jamie F
I'm using a single 35A rectifier for 10 LM3886's and it generates no appreciable heat. |
At full power? At 1 amp DC per channel, that bridge has got to be dissipating at least 15 watts - it should be heating up.
JoeBob, what about separate power supplies for 3 channels / 2 channels? Or is it too late?
Using just the +ve or just the -ve from a bridge is just a full-wave centre-tap connection (only two diodes are used). See the ESP article on power supplies for information on this. |
|
|
| JoeBob |
| Well, I've only got 2 caps and two transformers. And if I wanted to use multple power suplies less power would go to some channels then others, which I'd rather avoid. I'll just try using a bridge for the positive and one for the negative and see how it goes. When only using either the positive or the negative poles of the bridge, the opposite pole is sent to ground, correct? |
|
|
| paulb |
| JoeBob, could you let us know what your transformer voltages are and what DC voltage you expect to get? I'm a bit confused. The arrangement you mentioned doesn't use the centre-tap of the transformer, but I was talking about using it as ground. Your way will get twice the DC voltage. |
|
|
| arnach |
People seem to be saying that it's ok..
If you pull 30A amps out of a 35A bridge recitfier it will BURN UP. I've seen it happen many times.
Also, why don't you at least spec a 50A amp part or use individual diodes? Using individual 60A diodes you'd be fine..
-- Aaron |
|
|
| blmn |
Arnach,
If you use correct sized heatsink the bridge will not blow up considering 85% of their continuous nominal capacity. I had some in my bench power supplies with this problem when I was testing some class B amps. I solved it increasing the heatsink. However, there are some manufacturers wich use peak, or not continuous, current value information for their devices as a basic information, and, in these these devices have much less continuous current capacity.
regards |
|
|
| JoeBob |
Well, I checked and the peak for my bridge is alot higher then 35A, in the hundreds, although I can't remember exactly what.
paulb, sorry, I don't really understand how you mean to connect it. Each transformer is a 750VA +/- 24VAC toroidal with dual primaries and dual secondaries. The DC voltage I'm gathering I'd get is probably around 28VDC or 30VDC, it doesn't really matter exactly, it's for some LM3886's. I hope that clarifies some things, sorry for my ignorance by the way, I just don't see how I can use one transformer for the + and one for the -, if using them the center-tapped way, wouldn't that produce twice the DC voltage, compared to using them in parallel? |
|
|
| AudioFreak |
Joebob,
i'd wire the primaries as per your mains voltage then wire the same secondaries of each transformer together and use 2 bridges to form the rails .... the + of 1 bridge goes 2 +ve rail, -ve goes to gnd, +ve of other bridge goes to gnd, -ve of same bridge goes to -ve rail.....
i'll post a pick of how this looks later if you want. you've just got to make sure all the AC phasing is correct. |
|
|
| blmn |
JoeBob,
Maybe something like this was the idea of PaulB. Don't worry about the polarities of the transformers.
regards |
|
|
| blmn |
| OOOPS, the capacitor at the negative side is inverted. Sorry... |
|
|
| JoeBob |
| Oh, now I understand. Thanks, I'll give that a shot. Thanks for all the help guys. |
|
|
| AudioFreak |
| or you could try this....... |
|
|
| JoeBob |
Thanks AudioFreak that diagram's really helpfull. I'll go that route, I'll feel alot safer then just using one 35A bridge.
As for why I didn't just go buy a 50A, because the only ones at my local shop was 35A and lower, and the 50A is like 5 times the price of the 35A if they order it. So I find this solution alot better.
Thanks again for all the help guys. |
|
|
| paulb |
Sorry, I was out last night and didn't get to answer your question. Blmn's diagram shows what I was suggesting.
Audiofreak's diagram should work fine too, I think. |
|
|
| Helix |
i think audiofreaks' hook-up is better then blmn's 'cos you cannot garantee that the two transformers are exactly the same voltage and hence have slightly uneven rails. Even tho' this doesn't matter much is just better design.
BUT make sure the tansformers are the right way round with respect with each other. Check the AC voltage between taps with DMM, if 0 volts then they are in-phase, if the expected voltage then they are 90 degree out out phase. Just a thought:p |
|
|
| paulb |
| quote: | Originally posted by Helix
i think audiofreaks' hook-up is better then blmn's 'cos you cannot garantee that the two transformers are exactly the same voltage |
Um, audiofreak's hookup wires two transformer secondaries directly together - the consequences of uneven secondary voltage would be much more severe in this case.
I think JoeBob is using two identical transformers; if not, I wouldn't use audiofreak's arrangement. Even if they are, I highly recommend fusing each transformer separately.
I totally agree with the comments about checking carefully for the right phasing of the transformers. |
|
|
| Helix |
| quote: | | Um, audiofreak's hookup wires two transformer secondaries directly together - the consequences of uneven secondary voltage would be much more severe in this case. |
no, it's more lightly that the secondaries are more closely matched with each other then two diffrent transformers, in which case, if one is slightly lower from the other, the output voltage is the mid point.
HOWEVER how about using 4 bridges:eek: that way there is less current in each bridge, if thats the problem |
|
|
| Jamie F |
PaulB
One bridge for 10x 3886 - at full power? No, in normal use. The 10 channels aren't full range. I made the point to JoeBob about this because I think you've got to be going some to stretch the bridge. I accept that 2 is probably a better idea.
Jamie |
|
|
| blmn |
Audiofreak´s suggestion is better because filtering at the output is easier than Paul´s one, but its more complicated and all the cares related in this thread will be necessary.
regards |
|
|
| Helix |
actually i ment 180 degree out of phase, above. sorry
i think u know what i mean anyway |
|
|
| Geoff |
Time to put in my two-penn'orth. Both the previous proposals have problems. The first circuit does not fully utilise the transformer secondary and so potential VA capability is being lost. The second circuit requires closely matched secondary windings (< 0.5V difference) to avoid a high current circulating in the two parallel windings causing a heat build up and reduction in usable VA.
May I suggest that, for the cost of two additional 35A bridge rectifiers, the following circuit be used which will avoid these problems. This circuit has the added advantage the diode current will be halved. I am assuming that I have read this thread correctly and that you have dual-secondary, not centre-tapped, transformers
Geoff |
|
|
| subwo1 |
Since Paulb's and Audiofreak's diagrams show full wave rectification, the ripple is 120hz, and the filtering effects will be virtually identical.
Geoff's way is a good way, providing full wave rectification also. I would connect one bridge output from the top tranformer to the bottom capacitor, and one from the bottom to the top. |
|
|
| paulb |
| quote: | Originally posted by Geoff
The first circuit does not fully utilise the transformer secondary and so potential VA capability is being lost. |
Well, the transformer's VA capability is likely based more on heating effects due to losses than anything else; a full-wave CT arrangement only draws current on alternating half-cycles but you can pull more current on each pulse because the heating is based on average, not instantaneous current. However, there is effectively more winding resistance involved so you would indeed lose some VA capability.
Of the three I like Geoff's proposal the best if you can afford the 4 bridge rectifiers. |
|
|
| Geoff |
Paul
The utilisation factor for a transformer/rectifier circuit (defined as the ratio of the dc output power to the transformer rating required by the primary and/or secondary) is 0.572 for a full-wave centre-tapped circuit and 0.812 for a bridge rectifier circuit.
This means that either a smaller transformer can be used with a bridge rectifier or, for a given VA rating, more dc output power is available.
The foregoing figures apply to rectifier circuits providing a single polarity output (as was being done in blmn's circuit). When a centre-tapped arrangement is used to provide dual supply rails, the utilisation factor increases to that of the bridge rectifier.
Geoff |
|
|
| blmn |
Subwo1,
If the transformers were perfect you would be right, but filtering needs are defined by load, transformer regulation and frequency. Since the transformers and the load are defined (JoeBob's case), for the same transformer, for a given VA rating, as Geoff said, you have more power available and, if I don't miss something here, better regulation (less core losses, less wire losses etc).
regards |
|
|
| JoeBob |
I can see how Geoffs diagram works, and 2 more 35A bridges doesn't cost more (it's just 50A and above are ALOT more expensive for some strange reason). And if ti's the most efficient way I think I'll go that way. The PSU box is almost done so I'll wire it up tomorrow.
And Geoff, you're correct, two windings, not center tapped... |
|
|
| AudioFreak |
| Yeah, I agree with Geoff. His design probable contains the least number of potholes so i'd got with that. |
|
|
| paulb |
Geoff,
Do you have figures on efficiency as well? Utilization (oops I mean utilisation) you've defined as power out / VA rating required, but I think this might be different because of the pulsating nature of the current draw. What would the actual losses in the transformer be for the two cases?
My thinking is that by utilizing less power from the transformer it keeps its internal heating down. I normally like to run a transformer well under its VA rating.
Anyway, it's a moot point: JoeBob has the cash for the extra rectifiers; we all agree it's the best approach. Go for it, JoeBob!
Paul |
|
|
| Geoff |
The title of this post was 'Simple Bridge Rectifier Question'. Well, it wasn't so simple after all, was it? :-)
Geoff |
|
|
| Geoff |
Paul
Your last post came in whilst I was composing my previous comment. I agree with you about oversizing transformers, or putting it your way 'running them well under their VA rating'. I normally recommend a VA rating of 5 times the the amplifier rms power (in watts) for Class-A amps. Class-AB can be somewhat less.
I was quoting the Motorola (now ONSemi) 'Power Supply Design Manual', an essential read, though many other sources make the same point. The figures are calculated from the relationship between the rms and average current and so take into account the pulsating nature of the current draw.
You ask if I have figures for efficiency as well. How are you defining this if it isn't 'power in to power out' i.e. the utilisation factor?
Geoff |
|
|
| paulb |
Hi Geoff,
Yes, I was enjoying the irony of this thread's title as well.
I asked because power is not the same as VA. I'm curious as to what happens with the "extra" VA capability. I know that with a power factor other than unity, amperage is wasted in "imaginary power" that doesn't actually appear at the load; I'm wondering whether the extra VA capability required for a fullwave CT (for the same load) actually results in extra heat in the transformer - i.e. losses.
Just rambling, I haven't actually thought this through any further. I really do enjoy these discussions, there are a lot of very knowledgable people here and I always come away from these learning something new. Thanks for the reference on the ONsemi manual, I may pick up a copy.
Paul |
|
|
| subwo1 |
blmn,
Actually it does make sense that there would be a difference. I could see loading the same winding for a whole cycle instead of a half cycle could help utilize core flux and keep the windings more compact.;) |
|
|
| Geoff |
Paul
I wasn't quite accurate in my last post and managed to get the document title mixed up with another one I have read in the past.
The ONSemi document is No KB214/D entitled 'Rectifier Applications Handbook'. Be warned, it is a large file (336 pages, 2.8MB) so takes a while to download. Chapters 5 and 7 are the most relevant and it is the only source on the web that I have found that contains the full Schade diagrams for accurately calculating power supploies with a capacitor input filter. There is also a good section on choke input filters.
Geoff |
|
|
| AudioFreak |
| Geoff, just a small correction, the above handbook is 28MB |
|
|
| Geoff |
| Sorry, my mistake. That's two in one thread, it's just not good enough :-) |
|
|
| AudioFreak |
| quote: | Originally posted by Geoff
Sorry, my mistake. That's two in one thread, it's just not good enough :-) |
dont get too concerned .... i do it all the time. :P |
|
|
| Geoff |
H.H.
I always check that documents are still available before posting a reference to them.
A search for HB214/D in the technical documents section of the ONSemi site brings up the required handbook which can be downloaded as a pdf file:
http://www.onsemi.com/pub/Collateral/HB214-D.PDF
Geoff |
|
|
| HarryHaller |
Thanks for the actual link. I dug around on the site and didn't find it before. That's what I get get for sleep surfing. I don't know what insomniacs did at night before the web! I will mention again that many companies have thier datasheets and technical data on CD ROM free for the asking as service to the bandwidth challenged. There are several other appnotes of interest on the site as well. I urge all to take a look and thanks Geoff for the resource
Good day,
P.S. You had KB214/D instead of HB214/D before so it is still 3....
H.H. |
|
|
| Geoff |
Actually, I'll admit to 4 since 'supploies' is hardly correct. However, I will plead mitigating circumstances on this occasion as I was rushing to a job interview and didn't have time to do my customary check before hitting the 'submit reply' button.
My apologies to anyone who has wasted time trying to find this document because of my faulty reference number. It's not even as if H and K are adjacent keys and I could blame finger error. I think my mind must have been on other things at the time.
Geoff |
|
|
| Helix |
you have to be careful because if you,| quote: | | accurately calculating power supploies with a capacitor input filter |
that's quite the inverse of:| quote: | | accurately calculating power supplies with a capacitor input filter |
are you looking for a job then geoff? where and what? |
|
|
| HarryHaller |
I am sorry for picking on Geoff... he gave us a great resource location! Good luck with the job interview. I am also "between gigs" as the musicians say. The slow down in the industry has gutted the local telecom market and I am in the Telecom capital of the USA and was working for one the companies that was riding it out the best. This is why I have some free time for audio at present.
H.H. |
|
|
| paulb |
Okay, I'm helping this thread diverge but I can't help but add that I'm out jobhunting as well. It's really cuttting into my amp-building time.
Harry, if you were with one that was "riding it out the best" it couldn't have been Nortel Networks. I was there for five-odd years.
Paul |
|
|
| Geoff |
We're way off topic here so I'll probably get another email quoting the relevant extract from my webpage on forum etiquette :-)
H.H.
There's no need to apologise, I didn't feel 'picked on'. I made a mistake or two and I'll hold my hand up to that. I can recommend setting up a website to pass the time whilst 'resting'. The only trouble is, it takes over your life with all the emails it generates and leaves little time for job hunting.
Helix
Where: Lancashire. What: virtually anything (the over 50s can't be choosy).
Geoff |
|
|
|
