Choke values and why?

Looking at different schematics for example slow-blow preamp and Mila-preamps and so… there are sometimes mentioned chokes with certain values and some are more or less difficult to buy.

My question is if there is rule-of-thumb concerning the values (Henries)… med current values are "clear" to me but what about the Henries? Is it like values of capacitors concerning ripple or…what? I know what it is used for… and seen in many guitar amps before.

Best regards

/John

The simple answer is L=E/I where L is henries, E volts and I milliamps.

It also depends on where the choke is located in the circuit. As an input filter the choke, naturally, has to meet the PSU's critical specifications under heavy load.

To get an idea how dear Henry corresponds to capacitance, I would download Duncan's PSU Designer II.

The higher the Henries, the more the choke will resist changes in current. So in a PI filter of the typical tube amp, the inductor will allow DC voltage to pass without a drop, but will attenuate fluctuations in voltage.

The simple answer is L=E/I where L is henries, E volts and I milliamps.

Click to expand...

The actual equation is L=E/ (dI/dt), where dI/dt is the rate of change of current. This is a big difference:  the voltage drop across an inductor is the Henries times the rate that the current is changing.

dmp said:

The higher the Henries, the more the choke will resist changes in current. So in a PI filter of the typical tube amp, the inductor will allow DC voltage to pass without a drop, but will attenuate fluctuations in voltage.

Click to expand...

True in the abstract, however real chokes made with real wires will have non-zero DC resistance, which means even at DC there will be a real DC voltage drop across the choke that is proportional to this resistance. 

Most chokes will define this on their datasheet along with the expected Henries and tolerances.  For example, a Hammond 157G specifies 30H inductance @ 40ma current, with a resistance of 595 ohms.  This means that at rated current, the choke will drop about 24V DC.

Ok …nice to some info going here.

My concern is that the specified schematic I have read says that there should be a 10h 200mA and when I read about a such on Hammond it reveals that it is a monster… a solid 5 pounder… and the schematic says …Two! so even before I start solder it weighs  at least five kilos of metal!

Or… is the schematic "overkilled"?

http://groupdiy.com/index.php?topic=13171.msg538074#msg538074

i didnt use choke in PSU.

> the specified schematic

That schematic is REALLY EXTRAVAGANT!!

*Three* One-Thousand uFd caps?? Look at any other 300V 200mA B+ supply, you see 40 or 80uFd and maybe just two for the power-stage feed (more down the preamp chain).

> Or… is the schematic "overkilled"?

Post-apocalyptic.

> it weighs  at least five kilos of metal!

So you might consider a less extravagant design?

First step: plagiarize a power supply from a commercial design. 300V is very common 6V6 voltage, a very popular tube. Two 6V6 will draw about 100mA, so find a four 6V6 plan or double-up a two 6V6 plan.

If you plagiarize a lot, you see some trends.

The first cap can be 1uFd per mA. A bit more for low-Volt (heaters). A bit less for high Volts (plates). So for 200mA plate feed I'd think 100uFd.

This will give around 5% ripple, or say 15V. Push-pull pentodes (with separate screen filter and NFB) can usually run this way quite cleanly. Study Fisher, Dynaco, other classic tube stereo hi-fis.

Some power stages do not reject buzz so well. Single-ended triode is especially bad (tho a 300V 200mA SE triode would be an extreme beast). Increasing the first cap reduces buzz but very slowly. At some point you add a filter stage. This can be R-C or L-C. This gives MUCH better use of your money and space.

Let's look at the example. Because of the 10X extravagant first cap, the ripple here may be under 2 Volts. Then a 100 Ohm 1,000uFd filter. What is the reactance of 1,000uFd at 100Hz (double-frequency in 50Hz lands)? 1.59 Ohms. Plus ESR and stray wire loss, call it 2 Ohms. The 100 Ohm resistor and 2 Ohm impedance cap will drop ripple to 2/12 or 1/6th. So we are around a third of a volt ripple.

Then it adds 10H another 1000u cap. 10H at 100Hz is 6280 Ohms impedance. So ripple is reduced by 2/6202 or a factor of 0.000,32, for *a tenth of a milliVolt ripple*.

I can't think of any 200mA 300V load which needs sub-milliVolt ripple.

A large tube console might need less at the first stages. But they are typically less than 10% of the total load. It is much more economical to use small filtters (22K and 20uFd) at these points than to filter the whole 200mA to that degree.

And as a matter of practical electronics: you will never get 3,100:1 ripple reduction in a single stage. All that current banging around in mere wires, the "negligible parasitics come to dominate the action. 10:1 to 100:1 in a single stage is much more realistic.

OTOH, some designers have 5 pound chokes and soup-can caps just laying around the shop.

Thank you PRR

That was my thought as well... that the schematic is very extravagant. I just got curious because I thought it was a bit "over-sized".

I will stick to "normal" sized caps =)

/John

What is that schematic for?

The 1K/25W "HR2" resistor on HT output does not make sense at all - at 320V it will dissipate 102W, and probably kill both your mains transformer and the heavy choke while doing so..

Jakob E.

Well I have been interested in the slow-blow preamp and actually built one… I am at the moment using your psu from the G9 and it works just fine but since I am prone to build another I could not avoid seeing this psu schematic.

Some people say that I will need around 300 v ht but it works just fine with the G9 245 v ht… that is why this schematic is discussed here =)

I guess that 300 v would give some more headroom but the slow-blow preamp has plenty already.

/John

Forgot this one =)

Related question: Would it then be bad practice that if a 10h @ 200ma choke was called out, to substitute a 10h @ 300ma choke? The DCR rating is within 15 ohms of each other.

Thank you.

How much current are you drawing from the HT line?

I guess something like 25mA or so - why then would you need a 200mA choke?

Jakob E.

Building a 670 clone and that was the original value for the choke.

"I guess something like 25mA or so - why then would you need a 200mA choke?"

Well frankly I do not say I need it… I saw the schematic and just felt it was more suitable for Tjernobyl than the simple preamp… since I am a newbie and I mostly learn from trial and error… just curious in this case =)

"Less is more" in many cases and it seems like it is the same here =)

gyraf said:

How much current are you drawing from the HT line?

I guess something like 25mA or so - why then would you need a 200mA choke?

Jakob E.

Click to expand...

Zero. The secondary windings shorting rectifiers as shown will draw infinite current, leaving nada for HT, filament or phantom. Dunno if fuse, transformer or rectifiers will blow up first...

> 1K/25W "HR2" resistor on HT output does not make sense

Very hard to keep track when the image is (for NO good reason) three times bigger than my screen.

However it *looks* like that 1K is relay-switched from the 12V supply. The 1K is only "live" when the power is OFF. It won't cook day and night. It will cook about 3 seconds after the 12V fades away (decaying exponentially).

_______________________________________
> The secondary windings shorting rectifiers as shown will draw infinite current

Oh, cool! I like things which start with a BANG! I totally missed that.

_______________________________________

> the slow-blow preamp and actually built one...

The SB uses about 6 or 7 ma per stage, 13mA per channel.

The power supply I found in the same ZIP seems to be 200mA. It could feed *sixteen* of these preamps.

It isn't critical about power supply voltage. 350V won't melt it, 250V won't seriously reduce maximum output, and it would run on 150V though increasingly "colored" at high studio levels.

Let's start with a 230V AC winding, fairly readily available. This can rectify-out to 230*1.4= 322V DC.

Assuming two channels, 25mA total, we might start with 25uFd. Actually on a small job needing clean power, I'd buy a 10-bag of 47uFd and start with one 47uFd.

I have some rules of thumb for preamp filtering but I'll assume you can do busy-work and pick a five cap design (leaving a half-bag of caps for another project). If you were making a million units you'd want to do a sharp study of cap count and cost.

Since we picked the first cap bigger than 1uFd/mA, the ripple should be under 5%, or under 15V.

We need to clean some more.

Five caps and a C-R-C-R-C-R-C-R-C filter needs four resistors.

Let's assume these resistors will drop from 320V to 250V, or 70 volt drop. At 25mA this is 70V/25mA= 2.8K Ohms. We need this in four chunks, so four 700 Ohm resistors.

The reactance of 47uFd at 100Hz is 35 Ohms.

One R-C section of 700 ohm resistor and 35 ohm cap reactance will reduce 100Hz ripple 35/735 or 1/20th.

Three such filters will reduce 100hz ripple by 1/(20*20*20) or 1/8,000.

15V ripple at first cap, reduced by 8000, is 2mV ripple output.

This may be clean enough; or we may need 100uFd caps (double the cost will give 8X the filtering).

The OVER-KILL design uses three $20 caps and a $30 choke.... over $100. 

The more appropriate design uses five $4 caps and some cheap resistors... more like $30.

(In fairness, the KILL design will support 16 channels, the right-size was scaled for one or two channels.)

Computer sim of C-R-C-R-C-R-C-R-C filter with 47ufd caps gives 4V p-p ripple at first cap, 247VDC with 7 *micro* Volts ripple at last cap, at 25mA load.

Required ripple is "less than input hiss". We don't hear 100Hz as well as we do hiss, but steady-tone always stands-out, and this is a safe pencil-guess.

Input grid hiss on a tube will be 1 to 10 microVolts. The B+ ripple does not get into the grid, but the plate, after tube gain. Gain in this plan is about 20 in the first tube, we need less than 20uV ripple in B+ to first stage.

7uV is less than 20uV. Design appears adequate.

Thanks again PRR

Your information about this really simplifies my approach on power supplies… thanks

/John