Voltage Doubler

[triode17]

I am trying to run 12 preamp tube heaters ( six dual triodes) in a phono pre from a 12.6v regulated source. So, I used a voltage doubler from the 6.3vac tap. to give me 18v to start with, fine.
My ques.is: I am bypassing the rectifiers and 7812 v. reg. with small caps, but is there any value in bypassing the doubler caps? They are 4700u each.

 

[Retrovert]

You bet. Always bypass for better performance.

Larger capacitors very poorly perform when it comes to filtering ripple current. Goes right through them.

The other issue is that the three-terminal regulators like the 7812 can be noisy. Lower-noise regulators exist.

 

[triode17]

You bet. Always bypass for better performance.

Larger capacitors very poorly perform when it comes to filtering ripple current. Goes right through them.

The other issue is that the three-terminal regulators like the 7812 can be noisy. Lower-noise regulators exist.

The noise with a combined load of two heaters and resistors is less than 10mv P-P so I am not worried. It's starting up I am concerned about. i think the v. reg. will go into current pulses until the tungsten resistance rises enough for it to lock in at 12.5v. Will try the test tonight. The TO-3 version will be easier to control the heat on.

 

[Retrovert]

The other issue is that any HF noise on the line will go right through. So if you have any SMPS producing hash it gets into the system. Decoupling plus some HF filters fixes that.

You might want to current limit the filament at startup, because otherwise it will be a massive pulse of initial current as the filament heats. That terribly stresses the filament; light bulbs fail when turned on for the same reason.

 

[petehall347]

ecc83 seem to light up when starting up . yes a bit random i know . ignore me . found some lovely single malt i have .

 

[arts]

Keep in mind that you will have to seriously derate the transformer seconday current (approx. 50% or more) when feeding a voltage doubler.Nothing in life is free

 

[triode17]

I did the test last night and i was able to light up all 12 heaters with one TO220 7812. What happens is the IC goes into short circuit mode which pushes 2.2A into the load. Then, it recovered and came up to 12.5v within 1 second. These are quite rugged IC's !! I didn't have to derate the power tx. as it produces 6.3v @ 2.5A or 12.6 v at 1.25 A. I only need 0.9A so all is good, but I will still use a TO-3 case to spread the heat out. The little TO-220 gets damn hot. For testing, I simply mounted it on to a piece of 0.040" alum. about 2.5" sq. Plus, i added a 0.1uF Ceramic cap across all the heaters.

 

[BinaryMike]

Your transformer is at risk. Due to rectifier 'form factor,' 0.9A DC load current requires about 3.5A AC from a secondary winding working into a doubler rectifier setup.

In addition, you probably don't have enough raw DC voltage headroom for regulation with a 7812. The slightest decay in AC line voltage will cause ripple spikes to appear in the regulated DC output. A low-dropout regulator chip will improve matters slightly.

 

[triode17]

Your transformer is at risk. Due to rectifier 'form factor,' 0.9A DC load current requires about 3.5A AC from a secondary winding working into a doubler rectifier setup.

In addition, you probably don't have enough raw DC voltage headroom for regulation with a 7812. The slightest decay in AC line voltage will cause ripple spikes to appear in the regulated DC output. A low-dropout regulator chip will improve matters slightly.

Let me explain further. I am using SB530 Shottkey rectifiers for low forward drop. The lowest voltage trough is 15.4v and the output is sitting at 12.3v, so I have 3 v. of headroom. The circuit regulates down to 112 Vin. I wouldn't record or play through it if the line was that low anyway.

 

[triode17]

You bet. Always bypass for better performance.

Larger capacitors very poorly perform when it comes to filtering ripple current. Goes right through them.

The other issue is that the three-terminal regulators like the 7812 can be noisy. Lower-noise regulators exist.

Would you recommend using the 'two decades down method' of adding a 47uF cap to a 4700uF cap? Or, simply use a 0.1uF?

 

[Retrovert]

Adding two decades down could certainly not do any harm, and can help with faster response to transient demands.

I think your biggest bang for the buck comes from adding relatively small values of 4.7 uF, 0.1 uF and 0.01 uF. That should remove any higher-frequency ripple and noise, and it drops the ESR/ESL of the large capacitor. This has the benefit of moving any self-resonance way into the MHz band.

As long as you're doing this it can be better to construct larger capacitors using smaller ones because the parallel construction drops the ESR/ESL of the bundle. The difference in price is very small. Yes, the packaging costs a small amount, but overall it's maybe a few dollars, certainly a fraction of the capacitor cost. Going up in size sometimes costs a bit more so that offsets the multiple unit issue.

Much of this matters far more for digital logic and converters, of course, which are very sensitive to noise, but a cleaner supply is always a good thing.

 

[BinaryMike]

Let me explain further. I am using SB530 Shottkey rectifiers for low forward drop. The lowest voltage trough is 15.4v and the output is sitting at 12.3v, so I have 3 v. of headroom. The circuit regulates down to 112 Vin. I wouldn't record or play through it if the line was that low anyway.

Your transformer is still at risk. See the second page of this doc on rectifier form factor from Stancor: https://www.mouser.com/ds/2/459/pg_006_7-1212584.pdf

 

[triode17]

Your transformer is still at risk. See the second page of this doc on rectifier form factor from Stancor: https://www.mouser.com/ds/2/459/pg_006_7-1212584.pdf

Please enlighten me about Form Factor. Here is the schematic so we are all "on the same page", pun intended.

 

[triode17]

Here's an explanation of Form Factor from the IEEE, dated 1915 !!! Form factor is significant in the study of transformer losses; as is well known, hysteresis loss is small when the form factor is large, and vice versa. Every wave shape has a definite value of form factor; but the converse is not true, for a particular value of form factor does not indicate a particular wave shape. A wave may contain a third harmonic equal to seventy five per cent of the fundamental and still have the same form factor as a true sine wave. Form factor, therefore, has no general significance as an indicator of wave form or wave distortion. A general expression for form factor is derived in terms of the relative amplitudes and phase positions of its harmonic components; curves are drawn showing the variation of form factor with the amplitude and phase of the third harmonic. Various wave forms are shown, very unlike in appearance, having the same form factor.
I don't see the problem. The form factor is high at 4, so wouldn't the losses be small?

 

[BinaryMike]

The "Form Factor" referenced in Stancor's literature is the amount by which DC load current must be multiplied to get effective AC current in the secondary winding, for each form of rectification. It can very somewhat, depending on filter capacitor size, ESR, etc. Hammond's numbers are a bit less conservative, perhaps because their transformers are slightly over-designed relative to specs. I generally adhere to Stancor's guidance unless I'm going to take the time to conduct a thorough simulation or measure actual operating temperature rise in the winding. True-RMS ammeters often read incorrectly in this situation because rectifier conduction spikes can exceed the meter's allowable crest factor.

 

[triode17]

Mike - I look at it this way, the transformer is rated for 43 VA, I'm using 36, so it's running at 83% and it only gets warm, after 90 minutes. I don't see the concern here. What would you do?

 

[BinaryMike]

Mike - I look at it this way, the transformer is rated for 43 VA, I'm using 36, so it's running at 83% and it only gets warm, after 90 minutes. I don't see the concern here. What would you do?

If we accept Stancor's advice without qualification, then you're running a 2.5A heater winding at 0.9A * 4 = 3.6A, which is 44% over spec for that winding. If I felt strongly committed despite the odds, then I would run a temperature rise test. You must use a four-wire ohmmeter for this, but ad-hoc solutions are available. Just measure winding DCR, run the power supply for 2~3 hours with all windings fully loaded, then measure winding DCR again. If it increases more than 20%, that means temp rise was in excess of 50 degrees C and you're not in safe territory. If you have detailed knowledge of the transformer's insulation system then greater temp rise may be acceptable, but voltage loss might preclude use of this transformer anyway.

 

[triode17]

If we accept Stancor's advice without qualification, then you're running a 2.5A heater winding at 0.9A * 4 = 3.6A, which is 44% over spec for that winding. If I felt strongly committed despite the odds, then I would run a temperature rise test. You must use a four-wire ohmmeter for this, but ad-hoc solutions are available. Just measure winding DCR, run the power supply for 2~3 hours with all windings fully loaded, then measure winding DCR again. If it increases more than 20%, that means temp rise was in excess of 50 degrees C and you're not in safe territory. If you have detailed knowledge of the transformer's insulation system then greater temp rise may be acceptable, but voltage loss might preclude use of this transformer anyway.

But Mike,Its working fine. The TX only gets warm, maybe 35 deg.C. on the laminations, why worry, I'm not.

 

[BinaryMike]

The TX only gets warm, maybe 35 deg.C. on the laminations, why worry, I'm not.

I'm merely offering the opportunity to perform due diligence. Understand that a single winding buried inside the transformer can have spot temperature far higher than transformer bulk temperature, and insulation failure on any winding is catastrophic. I just got through checking actual temperature rise in the heater winding of a salvaged transformer that I'm considering for use in a line amp project. Cold DCR was 0.36 ohms, hot DCR was 0.40 ohms, making 11% rise, which I consider satisfactory. I would have rejected it if the rise got near 20%.

 

[triode17]

I'm merely offering the opportunity to perform due diligence. Understand that a single winding buried inside the transformer can have spot temperature far higher than transformer bulk temperature, and insulation failure on any winding is catastrophic. I just got through checking actual temperature rise in the heater winding of a salvaged transformer that I'm considering for use in a line amp project. Cold DCR was 0.36 ohms, hot DCR was 0.40 ohms, making 11% rise, which I consider satisfactory. I would have rejected it if the rise got near 20%.

I do appreciate your concern Mike. I may try your DCR test just to be sure. Will answer back.