Bias Supply

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

I've drawn a simple Bias supply with a simple program, Fritzing. The caps are in series (MKP 0.1µF, 250VAC) because I have a lot of them left and I'm using a neon blink charger at the end, I just don't know what value the cap has to be that's parallel with the neon bulb ? Does anyone have a suggestion or is this value not that critical ?

Many thanks guys.


  • Bias Supply.png
    Bias Supply.png
    107.1 KB · Views: 194
The value of the capacitor in parallel with the neon bulb isn't critical, but it does determine the blink rate. For a given leakage or charge current, higher values of capacitance will blink more slowly then lower values. Typically, values between 0.1uF and 0.01uF are used. For reference, Quad uses 0.047uF.
I have to ask......what am I missing with out the neon bulb in my bias? .......dose this neon bulb do anything for the sound or bias...are just tell us that we have leakage.... lot are littel..... an as we know we all have leakage.... what well most do to fix ...if you get more blinks...
I can see this as a test setup maybe ...for the panels at first ck...or time to time..
I have had..... quad 57s an they played fine with minny bliks....thay were 50 years old stock panels......thanks for any an all info on getting better sound out of ESLs
Thanks for the reply, I've put a 0.1 uF Cap in that position because I have a lot of them in different sizes, is there a way to know or to calculate what the blinking rate should be ? Will a 0.1uF, 600V AC cap suffice ?

Also, the resistor value in series with this cap, let's say 24Mohm, is this too high or too Low if the coating is Licron Crystal ? Maybe That's a bit off topic but I was just wondering.

Thanks for the help Bolserst

the cap across the blinker/flashlight is charged up by the supply.
When the voltage across its pins reaches the ignition voltage of the blinker, the blinker fires and discharges the cap.
The charging process restarts when the the blinker 'switches off' again, at a voltage which typically isn't specified in the DS.
The ignition voltage of the flashlight is stated in its specs (94V is a common and practical value. Just ensure that the rating of the cap is higher).
And You know the capacitance of the cap.
Now the problem is that the charging current is also dependant on the charging resistor and the leakage.
There's a rule of thumb for the charging time: tau~5x RxC
As the associated tolerances of all three 'participants' are quite high the results would be just rough estimates instead of precise figures.
The blinker is a nice qualitative measuring device but a bad quantitative measuring device.
Use it as such and don't worry about 'precision'

Last edited:
Ok...right now I remember night I would play my 57s in the the flashing light would drive me nuts...I put tape over the red bulb...
All ESL panels leak I see the neon bulb .. dose nothing...but flash...
May turn some off....they cant fix the ESL panel or make it flash less...if some one tell them flashing to much...
thanks for your info an time...have fun
Just curious on how this works.....

So you put the neon light in series with a 10 MOhm resistor, in series with the output of the step up ladder (which will be a few thousand volts) and then put a .1 uF cap (or whatever) across it. So if you have no leakage (ideal) the light would just stay on constantly? If you do have leakage, the slower it blinks the more leakage you have?

Think I am missing something here :)

yes, You missed it ;)
At the beginning both, the output of the bias supply and the membrane are at 0V.
When the supply is switched on its output voltage rises.
As the cap blocks dc and the flashlight also doesn't conduct, the membrane remains at 0V.
Now the cap is charged up and the voltage across its pins rises too.
When the voltage reaches the ignition treshold of the flashlight it fires, thereby acting like a switch that is closed, it conducts.
The cap discharges through the flash into the membrane.
Its voltage sinks and the memrane voltage rises (membrane is charged up).
When the cap voltage has reached a lower limit the flashlight stops firing (like a switch that is opened) and the charging process begins again.
The only difference is that the membrane is now at a voktage >0V.
As while the start up the voltage of the supply rises quickly the firing rate of the flashlight is high.
The rate of flashes sinks with rising membrane voltage/charge.
When the membrane is charged up fully (exactly: to a voltage lower than the flashlights treshold voltage below the bias supply voltage) the blinking stops completely.
In praxis the blinking will never stop completely as there is always leakage (even if not connected to the membrane, but just hanging in free air).
Which means, that the membrane voltage sinks slowly and the voltage over the cap rises.
When the treshold is reached the flashlight fires thereby recharging the membrane ... and so on and so on.
A slow blinking rate of say 2/min indicates low leakage/fully charged membrane while 2/sec indicate either start up or heavy leakage.
The blinker acts like a switch, that decouples the membrane from the bias supply when its off and connects to the supply while it blinks.
As the cap and flashlight are bi-directional the blinking will also happen when the bias supply is switched off and its voltage sinks blow the membrane voltage.

Last edited:
The neon bulb also isolates the panel from the power supply. so until the bulb blinks, they are operating in constant charge mode. In reality, there is leakage through the air and supports. Also the high resistance series resistor is also effective to isolate the panels from the power supply, so IMHO, the neon bulb is just window dressing. I love them on the newer quads, because it's a great way for customers to prove that they have a problem with their panels. in the Acoustat 121 bias....the 500mg is out...? An the neon bulb setup goes in..right.....??
Any one think this setup well give better...or is just a diff way to.....isolate the panel.... from the power supply it noise??
Seeing how my ears tell me the bias is in the Audio path the type bias feeder res...can change the sound.....this neon bulb has to be setup in my ESLs....just to hear.....thanks


  • images[1].jpg
    7.8 KB · Views: 83 in the Acoustat 121 bias....the 500mg is out...? An the neon bulb setup goes in...
I would suggest leaving the 500Meg in at first and just add the (neon bulb/cap) in series between 500Meg and the panels. Try it and see what the blink rate is. Personally, I would always leave at least 10Meg in series with the neon bulb to limit charging current, especially if your panels are fairly leaky.

The neon bulb has low "ON" resistance(1K - 2K) when it flashes, and extremely high "OFF" resistance when it is unlit. Ideally, if your ESL has low leakage, the neon bulb rarely flashes and your ESL panel is, for all practical purposes, completely isolated from the HV supply.
is there a way to know or to calculate what the blinking rate should be ?
As Calvin mentioned, the neon blink rate provides a nice qualitative indication of leakage…in particular when a panel begins to develop increased leakage with age. For quantitative measurement of leakage currents, I have used a series string of 10K, 100K, & 1M resistors between the transformer CT and the HV supply. Leakage current is determined by measuring voltage drop across these resistances with a 10Meg input impedance DVM.


  • HV_Leakage_Current.png
    24.3 KB · Views: 82
Dose the neon bulb have to be used in this setup??...or just the res & cap work??
if as stokessd states in his post......the neon bulb is just window dressing...??
an were do most get the..what 90v neon bulb?...what other bulb would work..? RS has 110v &120v red ones
Last edited:
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.