Using a Zener Diode to Regulate Heaters?
Moderators: CurtissRobin, colossal, zaphod_phil, Daviedawg, Graydon
-
- Verbose poster
- Posts: 1676
- Joined: Thu 09/21/06 2:00 am
- Location: Upstate NY, USA
Using a Zener Diode to Regulate Heaters?
OK guys and gals, I've been checking out my amps, and ALL of the ones I've built with Hammond iron are running at least 10% over on the heaters. This makes sense as my wall voltage is typically between 125 and 130 volts.
So anyways, I know I can use a resistor to bring down my heaters, but using a zener has much more appeal to me since my heaters won't get under-powered if any of my builds go somewhere with a reasonable wall voltage.
So first I was looking at those cool 6.2 volt zeners from Digikey and realized that I don't actually have 6.3 volts anywhere, I should have 3.15 volts and negative 3.15 volts.
So I'm looking at the 3.1 volt zener diodes and thinking of putting one between each heater leg and ground, sort of in parallel with the 100 ohm resistors a lot of amps use. Now I'm thinking this won't work. Looking at just one diode, when it sees the 3.5 volts on it, it will effectively regulate it to 3.1, but when it sees the negative 3.5 volts, it will just short that to ground as the zener will no longer be reverse biased.
SO now I'm thinking I need a regular diode in series with the zener so that the negative swing will not get shorted... ok, this is starting to make sense, but then I will only have one zener operating at a time and I will need to double it's voltage drop. So basically I know that one side of my heaters will be at -3.5 volts, then I need the other to be at +2.8 volts, right?
Does this sound about right? Here is a schematic of what I'm thinking:
*The only thing I don't like is that I'm just dropping the heaters by a constant voltage, I would much rather have a 6.3 volt or 2 3.1 volt zeners in there to provide a constant voltage, not a constant voltage drop ...?
So anyways, I know I can use a resistor to bring down my heaters, but using a zener has much more appeal to me since my heaters won't get under-powered if any of my builds go somewhere with a reasonable wall voltage.
So first I was looking at those cool 6.2 volt zeners from Digikey and realized that I don't actually have 6.3 volts anywhere, I should have 3.15 volts and negative 3.15 volts.
So I'm looking at the 3.1 volt zener diodes and thinking of putting one between each heater leg and ground, sort of in parallel with the 100 ohm resistors a lot of amps use. Now I'm thinking this won't work. Looking at just one diode, when it sees the 3.5 volts on it, it will effectively regulate it to 3.1, but when it sees the negative 3.5 volts, it will just short that to ground as the zener will no longer be reverse biased.
SO now I'm thinking I need a regular diode in series with the zener so that the negative swing will not get shorted... ok, this is starting to make sense, but then I will only have one zener operating at a time and I will need to double it's voltage drop. So basically I know that one side of my heaters will be at -3.5 volts, then I need the other to be at +2.8 volts, right?
Does this sound about right? Here is a schematic of what I'm thinking:
*The only thing I don't like is that I'm just dropping the heaters by a constant voltage, I would much rather have a 6.3 volt or 2 3.1 volt zeners in there to provide a constant voltage, not a constant voltage drop ...?
0 x
- Dana-L
- Occasional poster
- Posts: 232
- Joined: Thu 04/06/06 2:00 am
- Location: The Pacific Wonderland
Hi Alexo,
I like your thinking on this. My wall power reads only 119 VAC or so but my heater supplies still put out 6.3 V or greater.
How about rectifying to DC, using the 6.2V Zener and adding an inrush current limiter while you're at it. Certainly much more complex but your valves will love you for it!
Cheers,
-Dana
I like your thinking on this. My wall power reads only 119 VAC or so but my heater supplies still put out 6.3 V or greater.
How about rectifying to DC, using the 6.2V Zener and adding an inrush current limiter while you're at it. Certainly much more complex but your valves will love you for it!
Cheers,
-Dana
0 x
-
- Occasional poster
- Posts: 433
- Joined: Wed 10/13/04 2:00 am
- Location: SF Bay area
I would just leave the heaters alone. 10% over shouldn't be an issue. Just looked at a RCA 12AX7 datasheet. The heaters are speced at +/- 10%.
Using a zener like you described is not regulating anything. All you would do is clip the tops of the sine wave. The AC voltage for the heaters is an RMS voltage. That means your peaks will be 1.414 times the value you are quoting. Based on your schematic, you will also have unequal AC voltages, this would just induce noise into the amp.
CSB
Using a zener like you described is not regulating anything. All you would do is clip the tops of the sine wave. The AC voltage for the heaters is an RMS voltage. That means your peaks will be 1.414 times the value you are quoting. Based on your schematic, you will also have unequal AC voltages, this would just induce noise into the amp.
CSB
0 x
-
- Verbose poster
- Posts: 1676
- Joined: Thu 09/21/06 2:00 am
- Location: Upstate NY, USA
ahhh well thanks guys, that does make sense. ...suppose I could put one on the center tap for the filaments?
The datasheets I've seen said something like "absolute maximum +10%" I know they're still working fine, but I've been lucky enough to come across a fair amount of Phillips, Mullard and Telefunken tubes and I'm inclined to baby them as much as possible.
The datasheets I've seen said something like "absolute maximum +10%" I know they're still working fine, but I've been lucky enough to come across a fair amount of Phillips, Mullard and Telefunken tubes and I'm inclined to baby them as much as possible.
0 x
-
- Occasional poster
- Posts: 347
- Joined: Wed 03/29/06 2:00 am
- Location: Adelaide, Australia
Your thinking is on the path to a shunt regulator. But with a few things missing. You can't regulate AC with a simple regulator - regulators only make sense regulating DC. AC by its nature is intrinsically varying. So you would buy into the whole issue of DC heaters first. A regulator for AC is in many respects equivalent to a power amplifier, both in concept and complexity. (Or alternatively there are ferro-resonant transformers that can regulate AC power.)
Any regulator must be designed to get rid of the power, so with a high current feed such as a heater supply you would be looking at getting rid of a few watts - which is a very beefy Zener.
If you really wanted to have a play the best bet would be to rectify the heater supply. Use a bridge rectifier, add heaps of filter capacitance, and then use a LM317 regulator to get the exact voltage. Regulator will need heat-sinking. I think the general concencus would be that there would not be enough gain to warrant the effort. However if no one ever bucks the trend and gives it a go we don't make so much progress
Any regulator must be designed to get rid of the power, so with a high current feed such as a heater supply you would be looking at getting rid of a few watts - which is a very beefy Zener.
If you really wanted to have a play the best bet would be to rectify the heater supply. Use a bridge rectifier, add heaps of filter capacitance, and then use a LM317 regulator to get the exact voltage. Regulator will need heat-sinking. I think the general concencus would be that there would not be enough gain to warrant the effort. However if no one ever bucks the trend and gives it a go we don't make so much progress
0 x
- Dana-L
- Occasional poster
- Posts: 232
- Joined: Thu 04/06/06 2:00 am
- Location: The Pacific Wonderland
Hi Francis,
Thanks for chiming in on this, I appreciate your engineering knowledge and always enjoy your posts.
I've got a couple of questions:
Regarding dissipation in a zener, wouldn't the series resistor between the rectifier diodes and the zener be the primary component where power is dissipated? (I realize that a diagram would help a lot, sorry!)
When rectifying the heater supply, is it necessary to filter out the ripple?
Cheers,
-Dana
Thanks for chiming in on this, I appreciate your engineering knowledge and always enjoy your posts.
I've got a couple of questions:
Regarding dissipation in a zener, wouldn't the series resistor between the rectifier diodes and the zener be the primary component where power is dissipated? (I realize that a diagram would help a lot, sorry!)
When rectifying the heater supply, is it necessary to filter out the ripple?
Cheers,
-Dana
0 x
- s2
- Senior Amp Tech
- Posts: 3182
- Joined: Thu 05/20/04 2:00 am
- Location: Dallas, TX
- Contact:
A typical schematic for DC heaters includes the rectifiers (normally FWB), a large filter cap such as 3300uF 16V, and a voltage regulator. If the regulator has a lot of ripple rejection, you can get by with a smaller cap, but I wouldn't leave it off completely--especially considering the tiny price.
0 x
-
- Occasional poster
- Posts: 462
- Joined: Mon 11/06/06 2:00 am
- Location: Shropshire, England
- Contact:
Zener noise?
You can get 3-pin fixed voltage regulators (5V, 12V etc) that require no setting up like the LM317, so only a rectifer, reservoir and smoothing cap are required. (And maybe an extra diode to raise the output voltage if you wanted to)
Anyway, on the subject of simple zener shunt-regulators, does anyone think they might be noisy enough to cause a problem in the heater circuit? I've been thinking about using them to regulate DC heaters on an up-and-coming project, but the noise issue worries me.
(Yeah I know the other problem is no-load conditions, but I think a fuse in series with the zener would take care of that).
Anyway, on the subject of simple zener shunt-regulators, does anyone think they might be noisy enough to cause a problem in the heater circuit? I've been thinking about using them to regulate DC heaters on an up-and-coming project, but the noise issue worries me.
(Yeah I know the other problem is no-load conditions, but I think a fuse in series with the zener would take care of that).
0 x
-
- Newbie
- Posts: 85
- Joined: Mon 01/15/07 2:00 am
I think the best way to get plane DC is using a N-Channel FET as voltage-follower. All you need is a FET that's rated at the current and power level it has to take and a reference voltage (needs to be smooth). The reference voltage Uref sets the output of the voltage-follower to a certain value (the output is independent to the input; as long as Uin > Uout). U ref is calculated like this:
Uref = Uout + Uth (threshold voltage; see datasheet of the FET)
It works out pretty well. However, I never tried this idea in a circuit that is rather sensitive to voltage rippels. But who needs that for heating?
Uref = Uout + Uth (threshold voltage; see datasheet of the FET)
It works out pretty well. However, I never tried this idea in a circuit that is rather sensitive to voltage rippels. But who needs that for heating?
0 x
-
- Occasional poster
- Posts: 347
- Joined: Wed 03/29/06 2:00 am
- Location: Adelaide, Australia
In working a lot of this out you need to look at where the power is going.
An LM317 is a series regulator - as are the 7805 7806 7812 and the like that Merlin mentioned. (And indeed a 7805 with a pair of diode drops in the negative leg will make a perfect 6.2 volt regulator).
All any regulator like this can do is regulate down - by dissipating some of the energy. So the voltage of the feed must always be greater than the desired output voltage, plus some margin for the internal device drops. So clearly rectified AC alone won't do - it drops to zero twice a cycle. Thus smoothing caps are needed - and the ripple at full load must ensure that the feed voltage stays above the needed feed voltage.
The regulator only dissipates the power due to the drop across it - not the power it feeds to the recipient. So a bridge rectified 6.3 volt feed is 8.9 volts. We want some margin over that to cope with bad mains so call it 10 volts. If we regulate down to 6.2 we have a voltage drop of 3.8 volts. A 15 watt dissipation is achieved with pretty close to 4 amps. However - an LM317 has a maximum forward current of 1.5 amps - so we can't use it. Frankly's suggestion of a FET voltage follower is a good one, and indeed hints at the usual answer.
For high currents one needs an external power device to do the hard work. But it is possible to use a three terminal regulator to control it - thus getting tight control and high current. Needs about a dozen components all up - so is more complex than where we started. Have a look at figure 14 here.
Then again, there is nothing to say that you couldn't implement per-tube regulation with an unregulated DC feed running around the amp. Then you could just use a heap of three terminal regulators. Use 7806 and you get solid 6.0 volts with no extra mess. It all gets a bit over the top
An LM317 is a series regulator - as are the 7805 7806 7812 and the like that Merlin mentioned. (And indeed a 7805 with a pair of diode drops in the negative leg will make a perfect 6.2 volt regulator).
All any regulator like this can do is regulate down - by dissipating some of the energy. So the voltage of the feed must always be greater than the desired output voltage, plus some margin for the internal device drops. So clearly rectified AC alone won't do - it drops to zero twice a cycle. Thus smoothing caps are needed - and the ripple at full load must ensure that the feed voltage stays above the needed feed voltage.
The regulator only dissipates the power due to the drop across it - not the power it feeds to the recipient. So a bridge rectified 6.3 volt feed is 8.9 volts. We want some margin over that to cope with bad mains so call it 10 volts. If we regulate down to 6.2 we have a voltage drop of 3.8 volts. A 15 watt dissipation is achieved with pretty close to 4 amps. However - an LM317 has a maximum forward current of 1.5 amps - so we can't use it. Frankly's suggestion of a FET voltage follower is a good one, and indeed hints at the usual answer.
For high currents one needs an external power device to do the hard work. But it is possible to use a three terminal regulator to control it - thus getting tight control and high current. Needs about a dozen components all up - so is more complex than where we started. Have a look at figure 14 here.
Then again, there is nothing to say that you couldn't implement per-tube regulation with an unregulated DC feed running around the amp. Then you could just use a heap of three terminal regulators. Use 7806 and you get solid 6.0 volts with no extra mess. It all gets a bit over the top
0 x
-
- Verbose poster
- Posts: 1676
- Joined: Thu 09/21/06 2:00 am
- Location: Upstate NY, USA
Well I gotta say a .25 ohm series resistor is looking a lot more attractive!
Thanks a lot for that explanation, Francis, it has greatly broadened my sparse understanding of regulators.
If someone wants to start a business, he (or she) could devise a way to build a high-current 6.3 volt regulator and fit it in something like one of those W* copper caps so you could just mount a socket, plug the thing in, hook your heater taps into it and pull a smooth and steady DC heater supply off of it. Wouldn't that be nifty?
Thanks a lot for that explanation, Francis, it has greatly broadened my sparse understanding of regulators.
If someone wants to start a business, he (or she) could devise a way to build a high-current 6.3 volt regulator and fit it in something like one of those W* copper caps so you could just mount a socket, plug the thing in, hook your heater taps into it and pull a smooth and steady DC heater supply off of it. Wouldn't that be nifty?
0 x
-
- Occasional poster
- Posts: 462
- Joined: Mon 11/06/06 2:00 am
- Location: Shropshire, England
- Contact:
-
- Occasional poster
- Posts: 462
- Joined: Mon 11/06/06 2:00 am
- Location: Shropshire, England
- Contact:
It's possible. Maybe with a try though, all the same. If you've got enough spare voltage to drop maybe you could put a pair of diodes in EACH leg, to drop 1.4V...Alexo wrote:Will I gain hum if my heaters are uneven? i.e. if one side is at 3.5VAC and the other 2.8, won't they not quite cancel/humbuck?
Maybe add a snubbing cap across the pair, too, if switching noise is a problem. Can't hurt to try it though?
0 x
-
- Occasional poster
- Posts: 225
- Joined: Wed 02/21/07 2:00 am
- Location: Silicon Valley
Yes they could on the pre-amp filaments, in theory, because they are center tapped internally. PA filaments are not center tapped.Alexo wrote:Beautiful! Thanks, Merlin!
Will I gain hum if my heaters are uneven? i.e. if one side is at 3.5VAC and the other 2.8, won't they not quite cancel/humbuck?
I suppose I could cobble up some kind of a diode based clipper circuit that would clip the top of the sine wave above some value, like 6.5 vac. But my preference is still leaning toward dc for pre-amp filaments, in which case adding a linear regualtor would pretty much abate the problem.
-g
0 x
Gary Moore
Moore Amplification
Moore Amplification
-
- Verbose poster
- Posts: 1676
- Joined: Thu 09/21/06 2:00 am
- Location: Upstate NY, USA
Thanks, Gary.
I agree that regulation is clearly the best way to do it, here I could drop the .7 volts and it would work great in my house, but it may be below 6.3 if I go somewhere else, etc., but I just don't think I'll be able to find space in the 5 amps I want to upgrade to insert all those regulating circuits. I suppose if I used a 200 ohm "hum adjust" pot instead of the 2 100 ohm resistors or the transformer center tap, that might help a little.
About the internal humbucking - yeah, that would theoritcally be an issue, then again, I've used 6SN/L7's a lot and others use EF86's and it hasn't been an issue for most folks, so I think I'll just give the diode thing a shot and see how it plays out in the real world, I'm betting (hoping) it will be fine,
Then again maybe I should just get one of
these
I agree that regulation is clearly the best way to do it, here I could drop the .7 volts and it would work great in my house, but it may be below 6.3 if I go somewhere else, etc., but I just don't think I'll be able to find space in the 5 amps I want to upgrade to insert all those regulating circuits. I suppose if I used a 200 ohm "hum adjust" pot instead of the 2 100 ohm resistors or the transformer center tap, that might help a little.
About the internal humbucking - yeah, that would theoritcally be an issue, then again, I've used 6SN/L7's a lot and others use EF86's and it hasn't been an issue for most folks, so I think I'll just give the diode thing a shot and see how it plays out in the real world, I'm betting (hoping) it will be fine,
Then again maybe I should just get one of
these
0 x
-
- Occasional poster
- Posts: 225
- Joined: Wed 02/21/07 2:00 am
- Location: Silicon Valley
If you can't find a 200 ohm pot, the 100 ohm pots are more common and will work just as well. Just make sure the pot itself is rated for at least 1/2 watt, and I think you should be fine.Alexo wrote:Thanks, Gary.
I suppose if I used a 200 ohm "hum adjust" pot instead of the 2 100 ohm resistors or the transformer center tap, that might help a little.
these
-g
0 x
-
- Occasional poster
- Posts: 462
- Joined: Mon 11/06/06 2:00 am
- Location: Shropshire, England
- Contact:
[quote="Alexo"]Thanks, Gary.
I agree that regulation is clearly the best way to do it, here I could drop the .7 volts and it would work great in my house, but it may be below 6.3 if I go somewhere else, etc.[quote]
You don't need to worry about that, you can run as low as 5V without any problems. You just don't want to run ABOVE 6.3V, unless you absolutely have to.
I agree that regulation is clearly the best way to do it, here I could drop the .7 volts and it would work great in my house, but it may be below 6.3 if I go somewhere else, etc.[quote]
You don't need to worry about that, you can run as low as 5V without any problems. You just don't want to run ABOVE 6.3V, unless you absolutely have to.
0 x