hello.
yes the heat dissipation could be a reason,e. g. two 12v/1w zdiodes in series give 2w............and a 24v/1w gives 1w.
another reason is the dynamic (=inner) resistance of the diodes.....values around 8v or so have the lowest dynamic res and this increases with higher z - voltages...................have a look at the datasheets.
greetings...............
yes the heat dissipation could be a reason,e. g. two 12v/1w zdiodes in series give 2w............and a 24v/1w gives 1w.
another reason is the dynamic (=inner) resistance of the diodes.....values around 8v or so have the lowest dynamic res and this increases with higher z - voltages...................have a look at the datasheets.
greetings...............
Zener diodes in the 4.7-5.6V range have a temperature coefficient near zero... a string of these will have very little voltage change as they warm up. If you use higher values, you can add a few forward biased ordinary diodes to the string to cancel the positive change with temperature. For example, an 8.2V Zener has about 4 mV / degC change with temperature. A forward biased diode is about -2 mV / degC. So two diodes plus the Zener will be about 9.6V and stable with temperature.
Re: and noise
Yup;
Highish voltage zeners are avalanche breakdown types with increasing Zr with voltage and produce horrific fine grass noise and have to be driven fairly hard to reduce noise. With their lousy performance I'm not the only one who doesn't like using them.
Best solution is as others mentioned, string LV (true) zeners; or I use several TL431 in series; for cremé de la cremé performance.
This reggy ic has never let me down.
richy
aardvarkash10 said:
Yup;
Highish voltage zeners are avalanche breakdown types with increasing Zr with voltage and produce horrific fine grass noise and have to be driven fairly hard to reduce noise. With their lousy performance I'm not the only one who doesn't like using them.
Best solution is as others mentioned, string LV (true) zeners; or I use several TL431 in series; for cremé de la cremé performance.
This reggy ic has never let me down.
richy
Jean-Paul_2 said:
Thanks for all the informative responses. Now that I have a better understanding, I'm wondering if my original idea still has merit.
Here's the deal. I have an old PP EL84 power amp that I like to fool around with. It has fixed bias and I want to ground the cathodes with a 1 ohm resistor and add an adjustable negative bias. The power transformer has unused 12.6 volt taps that I intend to bridge rectify and CRC filter to achieve a -13 volt bias. I want to adjust this bias between -13 and -10 volts. So, I thought I'd connect one end of a 400 ohm 10 turn pot and go to ground through a 10 volt zener, or perhaps a pair of 5 volt zeners as suggested. Actually there will be four of these pots in parallel. Of course the wiper goes to the grid as bias.
It still seems reasonable to me but I'm concerned about noise and a few data sheet specs I don't completely understand. The four parallel pots will act as 100 R and the voltage drop is 3 volt worst case so the zeners will need to accomodate 30ma of current. There is a current spec but I'm not sure if it is maximum allowable or an inherent current (like DCR or impedence).
This is getting over long but I think you folks can get the idea. I need to select an appropriate diode or find a better way if the noise issue is going to be a problem.
Thanks
CD
rtsang said:
I fully agree with the above comment, but even a linear voltage regulator IC from the 79xx series could add some noise and is not really needed. Most important: in fixed bias amplifiers it is NOT recommended to regulate the bias voltage if your B+ is not also regulated: If B+ goes up (due to mains AC supply variations) and the bias is regulated your plate current will rise. With an unregulated (but well filtered) bias supply you will get some stabilisation effect on the plate current because B+ and Ug1 (bias voltage) will rise (or decrease) at the same time. This is why bias supply regulation is rarely used in fixed bias amplifiers. Just stay with the well proven classic scheme (unregulated well filtered bias supply with resistive dividers and pot's) and forget about noisy zeners or tricky circuits. This is a good exemple where simpler means better...
Couldn't you just replace the fixed bias resistor with a pot? Use a 10K or 20K resistor to ground on the wiper for safety (in case the pot ever shorts out)? This would allow you to adjust the neg bias and still have some resistance to ground should the pot ever short circuit. I have done this with a couple tube amp's I have and it seems to work fine. Measure the bias across the 1 ohm resistor to ground with a volt meter.
You're risking censure by the pot police, friend. They're addictive, pots. However, the argument against them is essentially this: #1 - over time their set-value(s) are not reliable, and #2 - they can become remarkably noisy signal sources.
Some people flaunt the pot laws by using multiturn trimmer pots. Twisted, this addiction is. But by their construction, they tend not to be unreliable at keeping a set-value, nor with their sealed construction (and difficult twiddly set-point kinematics) are they prone to noise. Usually the pot-police just cast a shady eye, and look the other way with trimpots.
Myself, I like 'em, especially when getting a novel design up and going. You can use them to set a wide range of quiescent and full-signal settings without needing to change out a resistor or ten. Tube circuits in particular are way better than average at allowing constructive trimpot addiction to have beneficial results: removed of their little electron vaporizers, tubes become non-conductive, thus isolating the pots. A couple of alligator clips and a good meter gives you direct, uncomplicated readings.
LOL
GoatGuy
Zener noise is massively overexaggerated by hi-fi types.
Zener diodes generate mainly shot noise. They probably have some pink noise too, but it is usually negligible for audio. Avalanche diodes generate shot noise but also additional white noise, plus a weird sort of pink noise called 'multistate' noise. But even so, you'll never hear in in a power output stage. Even in a preamp you have to make some pretty egregious mistakes to suffer any really serious noise degredation by using avalanche diodes for normal purposes. And you can always kill the noise with a big shunt capacitor.
Zeners are so noisy that I can't hear any noise from the horns of klipsch speakers. On my PP pentode, 82V+82V+24V 5W zeners string for screen grid supply.
As for thermal issues, it settles in about 2 minutes and after that it's rock solid.
So what's wrong with zeners ? is it too simple ?
As for thermal issues, it settles in about 2 minutes and after that it's rock solid.
So what's wrong with zeners ? is it too simple ?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.