'E'-Choke

Mike H · Post by **Mike H** » Sat Oct 03, 2009 3:08 pm

Anyone seen this?

http://www.hificollective.co.uk/catalog ... -5691.html

Blurb about ~

http://www.tentlabs.com/Products/Tubeam ... n%20UK.pdf

Mike H · Post by **Mike H** » Thu Oct 15, 2009 5:57 pm

Sorry for "borrowing" these quotes from 26 10Y KR300BXLS thread

Max N wrote:I was looking at the electronic choke that TentLabs sell - would there be any merit in a filter using VBE and electronic choke?

Nick wrote:Not sure just what the solid state choke thing is (or at least how its doing it), maybe a gyrator, but I didnt think that could be used for power. I suspect its a current source like regulator. Like the cap multiplier I suspect its just a filter and not doing any of the energy storage that a real cap or choke will do, but I dont know.

Mike H wrote:
Nick wrote:Not sure just what the solid state choke thing is ... I suspect its a current source like regulator
FWIW, so do I

A regulator of some sort, but with "floating input" I.e. not tied to a specific Voltage reference. Which got me thinking I have to admit

Nick wrote:I guess it could have a AC coupled reference and regulate to that. Yep, I can see how that might work.

Well I was doing some more thinking and found some better pix of the module.

There's one active device on a heatsink, but I can't figure why it's got five resistors. There's a small diode or maybe Zener on there as well.

Be nice to see the bottom of the board as well but extensive Googling did not reveal any such.

.

Mike H · Post by **Mike H** » Thu Oct 15, 2009 6:02 pm

PS that's not a LM317 it's thicker than that. My best guess is TIP142 darlington or something like that. Or a FET-lington.

Might even be an IGBT !

.

Mike H · Post by **Mike H** » Fri Oct 16, 2009 1:01 pm

I was trying to write this last night but then my Internet connection went off, just like that

Aaanyway, I was thinking could use an audio power MOSFET, given it's graph (below) is spookily similar to a pentode ~ where operating near the bottom then its current remains practically constant with varying Vds, the varying Vds being of course the ripple on the reservoir capacitor.

.

Post by **Nick** » Fri Oct 16, 2009 1:22 pm

Ok, yes, a CCS would be good as long as there was a constant current through the amp, but it would fail in a bad way once the current demand varied. I can't see any reason why it has to be a two terminal device though, it is connected to ground, just because the block diagram doesn't show a connection, it doesn't mean its not there. I wonder if its something like the floating op amp that MJ had, with a power transistor on the end.

Mike H · Post by **Mike H** » Fri Oct 16, 2009 1:33 pm

Yes that's the sort of thing, and as you say can't be a pure CCS in case the load changes, bu-ut, as I was also thinking like there, a V-drop across two resistors, with a cap strapped across the lower one, the junction point biases the output device.

All it's got to do is present a high impedance to AC, the actual DC operating conditions can be variable. Well have to be

I shall do some more thinking.

.

Mike H · Post by **Mike H** » Fri Oct 16, 2009 6:16 pm

OK well after some considerable cogitation I came up with this idea: (slightly modified from a previously similar idea)

The theory is the V-drop across R1 is say 5 - 10% that across R2. R2 is a high value like 1 Megohms.

As the Voltage on 'C-RESERVOIR' rises, the FET charges 'C-FILTER'.

After that it behaves like a "self adjusting" CCS, which, on having reached an equilibrium, Vgs (which determines o/p current) is kept constant (or rejects AC) by the charge on C1.

Theoretically then it provides a steady flow to 'C-FILTER' and the circuit while "ignoring" the ripple on 'C-RESERVOIR'.

VDR1 is a voltage dependent resistor or even a Zener to prevent max. Vds being exceeded on start-up (forces the FET to conduct more).

Reverse protection is provided by the device's internal diode. Also includes back-to-back Zeners (not shown) between G and S.

.

Post by **Nick** » Fri Oct 16, 2009 6:33 pm

Hmm, in fact is R2 needed?

Mike H · Post by **Mike H** » Fri Oct 16, 2009 6:52 pm

Funny you should say that cause ack-churley my "first version" didn't have it.

.

Mike H · Post by **Mike H** » Fri Oct 16, 2009 7:00 pm

And, here it is:

.

pre65 · Post by **pre65** » Fri Oct 16, 2009 7:17 pm

So Mike - have you built one yet ?

Mike H · Post by **Mike H** » Fri Oct 16, 2009 7:37 pm

No not yet but worth a 'speriment

.

Mike H · Post by **Mike H** » Sat Oct 17, 2009 1:52 pm

Yes this refuses to lie down quietly

Version #3

The problem I've got with the previous one is getting a workable V-drop acrose the FET (workable from the point of view of accommodating the max. ripple wave form), bu-ut, if just use two resistors to bias the gate the total V-drop is the ratio of input resistor to bias resistor. As the Vgs is really very small, this ratio gets rediculously large at currents up to 500mA. This V-drop is of course subtracted from the available HT.

Ideally don't want much more that say a 20V loss, and to stay more or less constant, only option is to 'force' this with a Zener. Then, to make sure some reasonable current is flowing through the Zener, requires R1 to 0V.

Otherwise it's more or less as before.

.

Post by **Nick** » Sat Oct 17, 2009 3:38 pm

No suggesting that its in any way related to the e-choke, but at least in sim, this gives about 50dB reduction in ripple. Select the voltage of the zener to exceed the peak to peak ripple thats on the input.

Not built it yet, but can't see any reason why it wont play.

Post by **Nick** » Sat Oct 17, 2009 5:47 pm

Ignore R2 and R3, just connecting r5 to the bottom of the zener is just as good.

audio-talk

'E'-Choke

#1 'E'-Choke

#2

#3

#4

#5

#6

#7

#8

#9

#10

#11

#12

#13

#14

#15