I was reviewing my ole Audio Amateur collection for articles to help me on a Hafler D200 refurbishment. I found an article in the four/1987 issue, by Fernando Garcia Viesca, Inductance for Your Supply Design.
He offers up an alternative to the traditional choke filter design. In the traditional design, the choke is located AFTER the rectifier and must be have an air gap and be sized to address the substantial DC current which magnetizes the core.
Mr Viesca's alternative has the inductor BEFORE the rectifier (indeed, in his proposal, he has a rectifier in each leg of the transformer leads to the AC terminals of the bridge rectifier. He claims this substantially reduces the DC current rating needed, and he provides scope pictures and measurements showing reduced peak rectifier current at 60 cycles, reduced ripple and inrush current (taken after the bridge). He also said he measured a 16 degrees F reduction in power transformer temperature.
Does anyone have comments on this approach? Mr. Viesca acknowledges there is the weight and bulk issue, but he (like myself) uses the choke to reduce inrush, with side benefits of reducing ripple and power transformer temperature.
Unfortunately, I can't understand how to calculate the necessary inductor value (although it may be that there is no critical value). I am willing to send a photocopy to anyone who may be willing to review the article. Just PM me.
He offers up an alternative to the traditional choke filter design. In the traditional design, the choke is located AFTER the rectifier and must be have an air gap and be sized to address the substantial DC current which magnetizes the core.
Mr Viesca's alternative has the inductor BEFORE the rectifier (indeed, in his proposal, he has a rectifier in each leg of the transformer leads to the AC terminals of the bridge rectifier. He claims this substantially reduces the DC current rating needed, and he provides scope pictures and measurements showing reduced peak rectifier current at 60 cycles, reduced ripple and inrush current (taken after the bridge). He also said he measured a 16 degrees F reduction in power transformer temperature.
Does anyone have comments on this approach? Mr. Viesca acknowledges there is the weight and bulk issue, but he (like myself) uses the choke to reduce inrush, with side benefits of reducing ripple and power transformer temperature.
Unfortunately, I can't understand how to calculate the necessary inductor value (although it may be that there is no critical value). I am willing to send a photocopy to anyone who may be willing to review the article. Just PM me.
This may work, but the inductor probably also reduces the power available to the load. Anything that changes the charging pulse shape tends to do that. For stuff that doesn't approach max load it's not a problem.
As for the inductance value, as I understand more is always better, as long as it's not saturating and is within it's current rating. It can take a large choke to improve on the isolation capacitance of the trafo.
Also, the choke is going to add to the leakage inductance of the transformer, so you'll have to update your snubbers if you use them.
I haven't seen the article you refer to though, so I may be missing something.
As for the inductance value, as I understand more is always better, as long as it's not saturating and is within it's current rating. It can take a large choke to improve on the isolation capacitance of the trafo.
Also, the choke is going to add to the leakage inductance of the transformer, so you'll have to update your snubbers if you use them.
I haven't seen the article you refer to though, so I may be missing something.
For small values of L, placing before or after the bridge is practically equivalent.
When L becomes larger (=useful?), the final OP voltage will tend to Vaverage for the downstream configuration, and zero for the upstream one.
The core magnetizing remark is complete BS: it is not the fact that the current is DC or else that matters, but the instantaneous value of the current, which will not differ much in any case.
When L becomes larger (=useful?), the final OP voltage will tend to Vaverage for the downstream configuration, and zero for the upstream one.
The core magnetizing remark is complete BS: it is not the fact that the current is DC or else that matters, but the instantaneous value of the current, which will not differ much in any case.
Some designs use loosely coupled transformers. These transformers have a high leakage inductivity, which helps in filtering. No need for an extra L.
I'm guessing by 'loosely coupled' you mean something like a dual-bobbin with bobbins on opposite sides of the transformer?
Yes, often with an air gap.
The toroidal transformer is not useable. But you could add some windings on the core for the choke.
The toroidal transformer is not useable. But you could add some windings on the core for the choke.
A transformer with an air gap would have quite a low magnetising inductance, hence high magnetising current and therefor uneccessary I^2R heating of the primary.
Leakagge inductance has the advantage that is not saturable , but to get values that are high enough tonbe of use to smooth the charging pulses it has to be very bad (leaking) transformer.
I did once added extra turna to one xformer and used a huge airgapped coil to smooth out charging pulses to reduce rms current, after all transformers available output power is mostly loss limited in the windings.
Leakagge inductance has the advantage that is not saturable , but to get values that are high enough tonbe of use to smooth the charging pulses it has to be very bad (leaking) transformer.
I did once added extra turna to one xformer and used a huge airgapped coil to smooth out charging pulses to reduce rms current, after all transformers available output power is mostly loss limited in the windings.
In the article, Mr. Viesca used 7.0 mH choke. He also suggested that chokes be wound using a toroid core and provided a table for the wire gauge to be used.
As far as reducing the overall voltage, that is something I am seeking because I do not need 100W a channel output (40-60W is fine), and would rather reduce the voltage to help provide more margin and longevity to the amp.
- Status
- Not open for further replies.