Best type of transformer for a preamp.

[beppe61]

Dear Sirs,

maybe this question has been debated to nausea in some other Thread here.
In this case please accept my apologies and I would be extremely grateful if someone could redirect me to the relevant link.
Nevertheless I think the problem is quite fundamental.
I found two opinions of well regarded audio designers one saying the opposite of the other.
One designer states that the transformer to use in a power supply for a preamp must be toroidal.
The other one instead says that EI or C-core type should be preferred, because they filter better high frequency noise due to their characteristic limited bandwidth.
Looking inside high-end preamps (phono and line) I can see the two/three types used indifferently.

Opinions would be very welcome and appreciated.
Kind regards,

beppe

 

[pooge]

A lot depends on where the transformer is mounted. If in the same chassis as the electronics, a toroid would be better since it radiates less magnetic field outside of the transformer that could couple into your front end and cause hum. You can add filters to reduce line noise.

If you put the transformer in a separate chassis, the other cores are better at reducing line noise.

 

[ampmade]

The two sort of transformer are good.In this two case, you must filter the rectifier with slow capacitor value (10 nF to 220nF) to filter HF switch off noise.The greater HF noise come from the rectifier and not from the transformer, the transformer filtering a part of main supply and a residual noise can stay but not as high as the rectifier noise.
You must compare the builder datasheet of the components you want to buy, to see the difference between us.
I have build a lot of power supply with toroidal transformer and they are working pretty good.
I think it's a question of design.A good transformer with a poor designed filter make a poor power supply.
I think (but i'm not sure) that predominantly the EI core transformers have a better weight/power ratio and the magnetic field loss are different from the toroidal form.
The price is a little bigger for EI core transformer too.

 

[beppe61]

A lot depends on where the transformer is mounted.
If in the same chassis as the electronics, a toroid would be better since it radiates less magnetic field outside of the transformer that could couple into your front end and cause hum.
You can add filters to reduce line noise.
If you put the transformer in a separate chassis, the other cores are better at reducing line noise.

Thank you sincerely Mr. Pooge for the kind and valuable advice.
I have to think about the two boxes solution.
The preamp is a DIY Bride of Zen.
I would like to try some different solutions for the power supply (single rail, 60V), maybe even using a LM317.

Kind regards,

beppe

 

[beppe61]

The two sort of transformer are good.
In this two case, you must filter the rectifier with slow capacitor value (10 nF to 220nF) to filter HF switch off noise.
The greater HF noise come from the rectifier and not from the transformer, the transformer filtering a part of main supply and a residual noise can stay but not as high as the rectifier noise.
You must compare the builder datasheet of the components you want to buy, to see the difference between us.
I have build a lot of power supply with toroidal transformer and they are working pretty good.
I think it's a question of design.
A good transformer with a poor designed filter make a poor power supply.
I think (but i'm not sure) that predominantly the EI core transformers have a better weight/power ratio and the magnetic field loss are different from the toroidal form.
The price is a little bigger for EI core transformer too.

Thank you sincerely for the precious advice.
HF switch-off noise from the rectifier is an issue I did not take into account !
I read about Scottky diodes not presenting this side effect.
I will read about that.
Thank you very much again.
Kind regards,

beppe

 

[richie00boy]

Schottky diodes do not suffer from turn-off spikes Beppe, do not worry about needing to use snubbers with these devices.

As for transformers, yes an EI will produce a cleaner supply, but at the expense of magnetic field leaking. So if you have a large case and can space the transformer away from circuits and wires, then use it. Otherwise use a toroidal and some extra filtering on the power supply.

 

[beppe61]

1) Schottky diodes do not suffer from turn-off spikes Beppe, do not worry about needing to use snubbers with these devices.
2) As for transformers, yes an EI will produce a cleaner supply, but at the expense of magnetic field leaking.
So if you have a large case and can space the transformer away from circuits and wires, then use it.
3) Otherwise use a toroidal and some extra filtering on the power supply.

Thank you very much Sir for your always kind, friendly and valuable advice.
1) I wonder if the effect of spikes is really audible in a line preamp. Actually Nelson Pass reccomends simple off-the-shelf generic diodes (IN4001 I believe).
With digital equipment maybe the gain in quality is more evident.
Digital do not like spikes IMHO.
2) I am leaning towards a solution with a separate case for the transformer alone and an umbilical power cord with the diodes and primary caps in the preamp box.
So a EI could be the best solution.
3) The fact that a toroidal would require extra filtering makes my task more complex.
I think that, as a start, it would be better to stay as basic as possible.
Nevertheless I am convinced that the Bride of Zen even in the original single ended version has some potential to sound really good.
I am truly optimist.

Thank you very much indeed.
Kind regards,

beppe

 

[richie00boy]

Personally I am not a fan of umbilical cords. The rectifier and capacitor charging currents are short periods and con create raidated harmonics. The last thing you want to do is radiate these out all over the place with an antenna, which is what your umbilical cord will be.

Really, the extra filtering for a toroid is minimal and not difficult to implement. In fact, if you are implementing a regulated PSU the filtering is very easy to accomodate as you can deal with input voltage fluctuations much better, so it's easy to add a simple filter to the front end of the regulator.

 

[beppe61]

1) Personally I am not a fan of umbilical cords.
The rectifier and capacitor charging currents are short periods and con create raidated harmonics.
The last thing you want to do is radiate these out all over the place with an antenna, which is what your umbilical cord will be.
2) Really, the extra filtering for a toroid is minimal and not difficult to implement.
In fact, if you are implementing a regulated PSU the filtering is very easy to accomodate as you can deal with input voltage fluctuations much better, so it's easy to add a simple filter to the front end of the regulator.

Dear Sir,
thank you again for your kind and very precious advice.
1) Actually your words about the better noise rejection of EI type transformer have convinced me at the point that I would be ready to use a shielded power cord if this can avoid the "antenna" effect.
2) Not difficult not for me. Anything turns to be difficult with my limited knowledge.
I have thought about inserting some mH in series with the voltage rails but I do not know how to calculate the optimum value and where place them.
Actually I would like to build a regulated single rail (60V) PSU.

Thank you very much indeed.
Kind regards,

beppe

 

[richie00boy]

Shielded cable will help a lot but still the underlying problem is there and may be worse than the hum pickup problem should you use all in one case.

The filter really is simple - following the rectifier and smoothing cap, just a 10 ohm 2 watt resistor in series with the rail with cap of minimum 470uF to 0V would do the job brilliantly.

 

[pooge]

The bottom line is to get a good transformer, no matter which kind. E-I cores can tend to make mechanical noise from vibration of laminations, more often than toroids. Transformers can also make noise if there's DC on the input. If you go with an external case, split your caps up into each case. This will keep your charge pulses out of the cord, and any resistance or inductance in the cord can help filter the second bank.

 

[AndrewT]

Hi,

I am leaning towards a solution with a separate case for the transformer alone and an umbilical power cord with the diodes and primary caps in the preamp box.

I would put the rectifier and first stage smoothing with the transformer, keeping the loops around the current routes as short and compact as physically possible.

You can add second stage smoothing at the preamp or even on board the preamp PCB (decoupling). This two stage smoothing gives a very effective RCRC PSU which when designed appropriately gives low levels of ripple at the pre-amp.

Pooge is telling you the same thing.

If you go with an external case, split your caps up into each case. This will keep your charge pulses out of the cord, and any resistance or inductance in the cord can help filter the second bank.

Be very careful with grounding, both safety and audio, when you use an umbilical.

 

[beppe61]

Shielded cable will help a lot but still the underlying problem is there and may be worse than the hum pickup problem should you use all in one case.
The filter really is simple - following the rectifier and smoothing cap, just a 10 ohm 2 watt resistor in series with the rail with cap of minimum 470uF to 0V would do the job brilliantly.

Thank you very much again Sir for your very helpful advice.
I will build that very simple filter.
I understand that in any case it should be very effective.

Kind regards,

beppe

 

[beppe61]

The bottom line is to get a good transformer, no matter which kind. E-I cores can tend to make mechanical noise from vibration of laminations, more often than toroids.
Transformers can also make noise if there's DC on the input.
If you go with an external case, split your caps up into each case. This will keep your charge pulses out of the cord, and any resistance or inductance in the cord can help filter the second bank.

Thank you Sir for the kind advice.
I will follow your suggestion.

Best regards,

beppe

 

[beppe61]

Hi,
I would put the rectifier and first stage smoothing with the transformer, keeping the loops around the current routes as short and compact as physically possible.
You can add second stage smoothing at the preamp or even on board the preamp PCB (decoupling).
This two stage smoothing gives a very effective RCRC PSU which when designed appropriately gives low levels of ripple at the pre-amp.
Pooge is telling you the same thing.
Be very careful with grounding, both safety and audio, when you use an umbilical.

Dear Mr. Andrew,
thank you for your kind and helpful instructions.
As you can very well understand I am just a beginner in DIY.😀

Nevertheless I have another maybe shameless question:
do you think that a LM317 regulated PSU has no chance to sound good in a line preamp ?
I must confess that I built a very basic prototype of 0-60V regulated PSU using a LM317.
Not only it worked but it sounded very impressive, of course IMHO. 🙄
Maybe this could mean that my whole system is not resolving enough to judge adequately difference in sound.
But I preferred the BOZ with the LM317 PSU to a my other old Bryston .4B.
And I was quite surprised by that.

Thank you very much and best regards,

beppe

 

[richie00boy]

The LM317 is a nice little chip and with the right implementation can perform really well. The only problem is that if you are have input voltage of more than about 40V it can present problems when the load capacitors are discharged at start-up.

Back to the umbilical cord. What others have suggested about putting the rectifier and first stage of caps in the box close to the transformer seems like a good idea, but still great care must be taken as you have the potential to have a star ground that is 'stretched out' over too long a distance.

 

[AndrewT]

Hi,
keep the audio star ground in the pre-amp case.

The PSU common links through the umbilical to the audio ground. This keeps the charging pulses and the resulting voltage pulses out of the audio ground.

The same philosophy applies if the PSU is close to the amp. Link the PSU common to the audio ground. Do not put the audio ground on the PSU common.

 

[beppe61]

1) The LM317 is a nice little chip and with the right implementation can perform really well.
2) The only problem is that if you are have input voltage of more than about 40V it can present problems when the load capacitors are discharged at start-up.
3) Back to the umbilical cord. What others have suggested about putting the rectifier and first stage of caps in the box close to the transformer seems like a good idea, but still great care must be taken as you have the potential to have a star ground that is 'stretched out' over too long a distance.

Thank you Sir for the reply.
1) this is a very welcome new, because the LM317 is so cheap and quite easy to use providing that Vin-Vout < 40.
It can be used even with high voltage at the input.
2) in the schematic I am referring to I have read that a diode protects the LM317.
3) I understand that an umbilical cord is a very unfortunate solution. Well, I will not use it then.

Thank you so much again and kend regards,

beppe

 

[beppe61]

Hi,
keep the audio star ground in the pre-amp case.
The PSU common links through the umbilical to the audio ground.
This keeps the charging pulses and the resulting voltage pulses out of the audio ground.
The same philosophy applies if the PSU is close to the amp.
Link the PSU common to the audio ground.
Do not put the audio ground on the PSU common.

Thank you very much for your kind and precious advice.
Kind regards,

beppe

 

[jackinnj]

Schottky diodes do not suffer from turn-off spikes Beppe, do not worry about needing to use snubbers with these devices.

as On-Semi says "...In circuits using Schottky diodes, the ringing energy can be high enough to generate a voltage higher than the diode breakdown voltage..."

this is more a problem in low level/high gain circuits where the radiated energy from a diode can cause all sorts of opamp havoc...and in digital circuits, etc.