-340Vdc bias from secondary taps (LONG)

[Tobacco Soup]

Hey folks,

I'm constructing an amplifier using a Hammond 278CX 400-0-400 power tranny, which does not provide a bias tap. The arrangement I employed was simple enough, but I'm getting WAY too much negative voltage for a bias cct.

First off, the first filter cap for the bias cct. (that will feed a TUT push-pull balance bias arrangement) is a 33uF/400Vdc cap because it was a nice small cap at a low value. This cap did see an over-voltage of some -440Vdc on the first attempt and wasn't changed yet (seems to check-out fine, and from what I've heard these caps have approx. 50V to spare from their rated capacity).

Next, I upped the value of the dropping resistors from 200k to 470k and now I get a slightly lower -340Vdc-ish bias supply voltage. What concerns me is that, even though it's still extremely high voltage, it still climbs steadily (about 1volt/2seconds). Something I am not accustomed to seeing in my limited experience.

Lastly, I thought my idea of a FW bias cct. was flawed so I removed one diode/dropping-Resistor 'leg' from the cct. Thinking that I just created a FWB arrangement with a very high (equivalent) internal resistance.


1st iteration: Full Wave scheme

400v ----> 220k---1N40007---33uF/400V positive side junction

400V ----> 220k---1N40007---33Uf/400V positive side junction


2nd iteration: Full wave scheme

400v ----> 470k---1N40007---33uF/400V positive side junction

400v ----> 470k---1N40007---33uF/400V positive side junction


3rd iteration: Half wave scheme

400v ----> 470k---1N40007---33uF/400V positive side junction

400v ----> open (as far as bias cct. is concerned)

Any help with this would be greatly appreciated! TIA!

[Enzo]

Well, first off you have not provided us with the schematic of what you made, or pointed us to whatever circuit you copied.

However, the voltage is not reduced merely by adding series resistance, the voltage is dropped by voltage division. it is the ratio of a series resistor with a resistor to ground that determines the voltage coming out.

If I read your description correctly, you have a series resistor from the AC to a rectifier and that to a filter. You need a resistor to ground - parallel to the cap essentially. Well, no essentially about it, that is where it belongs - across the cap.

[cbarrow7625]

What Enzo says is true. Another way to go about it is to use a series resistor & a zener shunt to get the voltage you want. The zener method should provide a little more stable bias voltage than the voltage divider (if that is what you think you need).

Either way, you are dropping a lot of voltage so make sure you carefully calculate the wattage ratings for those resistors or you will end up with NO BIAS if you fry a resistor.

Chris

[Satamax]

http://www.schematicheaven.com/ampegamps/v4bpowramp.pdf

[cbarrow7625]

Use a capacitor to reduce the 60Hz AC voltage! That's a great idea & really reduces the likelihood of burning up any resistors. Thanks Satamax.

[Tobacco Soup]

Thanks a lot guys for your rapid replies!

The power supply is pretty-much standard fare: a FW into a 'Pi Filter' for the B+, then the bias is (hopefully, properly) tapped from the 400Vac secondary(s).

The architecture of this build (something SLO-ish), where I use a separate PS board in a 'vertical' mounting arrangement for space conservation, left me with the decision to do all of the power rectification, splitting, what-have you, on the PS board along w/ the 'main' filter caps. The 'meat & potatoes' of the bias cct. is on the 'preamp' board close to the PI where I felt it belonged to save space.
Ultimately, the voltage dividing resistor is on the 'preamp' board with the rest of the 'meat & potatoes' (of the bias cct). Given the complexity of the build on the preamp-topology side of things, my logic is to proof each phase of the build as I go. I wrongly assumed that I could create something close to the required bias voltage before it reached the rest of the bias cct.

I had seen the ampeg schematic before, and it was an option I was/am willing to explore if I understood it better. Referring back to this ampeg schematic, is the 75k resistor after the reverse-biased diode the 'voltage divider' resistor that is functionally present in all the bias schemes that we're all familiar? TIA!


P.S. What is causing the slow build-up of voltage on this questionable bias cct. of mine? Faulty cap? Or, the lack of a voltage dividing resistor connected to ground and high value dropping resistors causing a slow climg to maximum possible potential voltage? TIA!

[Enzo]

If you don't have a complete circuit, there is no path for the charging current of the cap - as minimal as it might be. So the 10 Meg or so of your voltmeter impedance serves as the ground path. The cap is trying its best to charge up through that 10 Meg. It goes slowly.

An amp is a system, not a collection of parts all sitting side by side. So testing the supply in this fashion is like testing a joke a line at a time. Like this:

A horse and a mule walk into a bar.


How is it so far?

[Dave Curtis, dB AudioTech]

A horse and a mule walk into a bar.

So the horse says to the mule, "Why didn't you warn me?"

Sorry, couldn't help it.

[Enzo]

A horse walks into a bar and orders a drink. The bartender serves him a shot. Then the bartender says, "So, why the long face?"

[Tobacco Soup]

Thanks guys!

I know what you mean about individual ccts. vs. ccts. working in tandem. When I dissected the ccts. by function, I didn't get enough of the bias cct. into the 'suitcase' (bad metaphor )... I'll report back with my results as they become available.

"so the horse sees this Genie lamp sitting on the counter behind the bartender, and asks the bartender about it..."

[Tobacco Soup]

Thanks a lot for all of your help folks! Using Enzo's great suggestion, and referring to my texts, I now have a (almost) useful -92Vdc for feeding the bias supply. For those that would like a Lego-block/cut-n-paste solution to a similar scenario, here is what I used:

H.V. tap1 _____470k_____1N4007_____o__________o___________o_______ __> -92Vdc > rest of bias
                                                     |                 |                 |
                                                     |                 |                 |
                                                     |               150k          33µF/400Vdc
                                                     |                 |                 |
H.V. tap2 _____470k_____1N4007_____|                 |                 |
                                                                        |                 |
                                                                        |                 |
                                                                    ground       ground


Bear in mind, the resistor from "the rest of the bias cct." will be (functionally) in parallel with the 150k resistor shown above, and this resistor's value will decrease the voltage potential (if you accept that a -92V is a higher voltage potential than say -12V). Basically, if you need more negative voltage for your bias supply using this arrangement, you need to make the resistor to ground in "the rest of the bias cct." larger, if you need less negative bias voltage then that resistor to ground needs to be smaller.

I will edit as necessary to ensure completeness and correctness. Other than that, this is how one may implement a FW bias supply from a high voltage secondary... Regards,
TS

[Satamax]

Hi Tobaco soup.

Well, i don't get it, why do you use both the HV sides for your bias? And why don't you have a cap between thoses and the bias circuit. To me that doesn't look like a bias circuit but like a simple full wave rectifier with a voltage divider input.

[Tobacco Soup]

In lieu of a bias tap, the resulting -92Vdc goes where the bias tap WOULD go (less the extra diode). Hope that helps...

The "rest of the bias" IS where the other stuff is (resistor, caps, potentiometer) that makes a plain vanilla bias supply.

[Satamax]

I still don't get it, you don't have anything else connected to the HV tap1 and 2? If that's the case, Ok, but if not????

[Tobacco Soup]

I don't SHOW anything else, because the B+ PS cct. isn't pertinent to deriving a bias supply voltage from the high voltage secondaries... If you feel better about it, the B+ diodes are connected at the H.V.1 & H.V.2 taps and you can fill in the rest from there (FW rectification there too). The whole point to what I was endeavoring towards, was creating a useful negative voltage bias supply from the secondaries that deliver more than 4x what a 'usual' bias tap from a transformer would provide.

Furthermore, since this is going into a (hopefully) high-gain, high-saturation, type amplifier (i.e. noisy), I reckon a FW arrangement for the bias supply may actually help a little in the noise dept. Albeit, very little. In this case, every little bit should help. Besides, I'm not going to keep from sleeping because I used an extra diode and a surplus capacitor (which isn't necessary either).