DC Heater Circuit Question

[Ed_Chambley]

I got a circuit from a website Guitarnuts.com.  Attached is part of it and I used it in a Silverface Deluxe Reverb to quiet it down.  It had loads of floor noise and I just did this to experiment with the circuit with an amp with a few preamp tubes to see how it would work.  Well it worked fine.  I posted the entire schematic for the amp this came from as well.  It is a nice low watt high gain build.  I have not built the amp.


The amp builder stated he got around 5.8 vdc for the circuit and I did as well using a 6.3 with a bridge, but 6.3 x 1.41 is closer to 8.9vdc.  I cannot understand how it is possible from this circuit to get 5.8vdc as I do not see it possible for loading to do this.


Anyone have any idea how both of us would get the same reading.  He is using 2 preamp tubes and 2 power with a 6.3/2A heater winding and I am using 6 pre and 2 power tubes with 6.3/3A.


Anyone have any idea.  Here is his link and his description is under filament power supply.
http://www.guitarnuts.com/amps/nickelblues/index.php 


Anyway, I like the outcome, but would like to determine how we both ended up with the same voltage and I would also like to get the voltage in the 6vdc range.  It is simple and very quiet.  I plan to put it into the only high gain amp I have.

[kagliostro]

The amp builder stated he got around 5.8 vdc

May be he would write 8.5v and wrote 5.8v

in a low voltage like the heater voltage when you rectify, you must also remember that in the rectifier diodes or in the bridge there is a drop of voltage that usually is in the order of 0.7v

a friend that uses sometime DC heaters, if necessary (depends on the PT) uses an in series resistor to drop the voltage at the standard level

reading slowly the text, seems that he refers to the voltage he measured after the 1R resistor not before

K

[eleventeen]

You could be overloading the heater winding. When you first turn it on, with the tubes cold, do you get your 8.x volts and then it drops when the tubes light up?

[sluckey]

This is a quote from the guy's website...

... I used a simple pi filter with two 4700u caps and a 1R1 2 watt (actually, 5 watt was what I had handy) resistor. The power tube filament is supplied from the first stage of the filter and the preamp tubes are supplied by the last stage.

...This also reduces the filament voltage to the preamp tubes to right at the low edge of the nominal range (5.8VDC measured with mains voltage at 119VAC).

So, the 5.8Vdc he's talking about is AFTER the 1.1Ω dropping resistor. The voltage out of the bridge will be higher.

This circuit may be effective in a high gain amp such as he built, but IMO, is not very useful in low gain amp like a Deluxe Reverb or Vox AC-15.

[HotBluePlates]

The amp builder stated he got around 5.8 vdc for the circuit and I did as well using a 6.3 with a bridge, but 6.3 x 1.41 is closer to 8.9vdc.  I cannot understand how it is possible from this circuit to get 5.8vdc as I do not see it possible for loading to do this.

"... I do not see it possible for loading to do this."

Well, you know better than that. If the load were 0Ω, would that not reduce the output voltage below 8.9vdc?

We are used to looking at rectifiers for the high-voltage supply, their filter caps and the relatively small idle current. Here you have a relatively high current draw.

Kevin O'Connor pointed out that bridge-rectifying a heater winding and placing a nominal-value cap on it will yeild an output d.c. voltage near the number-value of the RMS a.c. voltage. Essentially, you have a water tank (cap) being filled by a hose at the top (the PT and rectifier) but with a hole at the bottom draining out water (the heater current draw). If the rate of fill (rectifier output current) of the tank is roughly-same as the rate of drainage (heater current draw), the water level doesn't rise.

We get used to thinking of current as a quantity, but it's really a rate: 1 ampere is 1 coulomb of electrons moving past a point per second.

Also remember you only get RMS volts * 1.414 as d.c. output voltage when current draw is near-zero. The problem is that the math for predicting the exact resulting d.c. voltage for a rectified a.c. voltage is pretty complex. If you want to read more about it, download RDH4 from the Library of Information (at the bottom of this page) and turn to Chapter 30. The process and data needed to get a result will give you a headache.

So, the 5.8Vdc he's talking about is AFTER the 1.1Ω dropping resistor. The voltage out of the bridge will be higher.

2 12A_7 tubes after the resistor, so 0.6A drawn. 5.8v + (0.6A * 1.1Ω) = 6.46vdc. Pretty close to what KOC predicted.

[Ed_Chambley]

HBP, Frankly I do know better.  I know enough to know loading can be very surprising.  The intention of this thread was not to apply it to the silverface Deluxe, it is just an amp I am needing to go through.  It needs a lot of work so I thought I would try this on the existing wires.  It worked well, but the amp is too valuable to leave this way.


I was surprised to find that no more than a bridge and a couple of caps would deliver a usable DC voltage since most of the circuits are elaborate and contain more than caps and resistors.  Just thought it interesting and very quiet as well and simply thought I would share findings.  I like simple things that work, but like I said I would like to see a little more voltage than 5.8vdc.

[HotBluePlates]

If you are looking to quiet down an amp using DC voltage, why not look at what the audiophiles do.   A lot of their designs use voltage regulators.    I wonder why a resistor is used to drop valuable volts.

Common voltage regulators are often "3v regulators" meaning they can keep a constant 3v (or whatever the rated voltage is) between their Adjust pin and the Output pin.

In order to regulate to 6.3vdc output, you generally need 11v+ of raw a.c. input, meaning you can't use a 6.3vac winding. The issue is the minimum voltage which much be dropped across the regulator plus the dropout voltage of the regulator plus the needed d.c. output.

The other huge issue is heater current. Regulator chips have a power rating which can be exceeded if you account for the current drawn from the heaters times the voltage drop across the regulator. Often you can only regulate some of the small tubes, and even then need to work on a current regulator which outputs a constant 300mA output (times the number of small tubes).

Or you have to up the regulated voltage to 12.6vdc so you're only supplying 150mA per small tube. Which again necessitates a higher voltage winding.

Regulated heaters are a pain, and that's why much ingenuity was spent finding ways to not need regulated heaters.

[smackoj]

<Regulated heaters are a pain, and that's why much ingenuity was spent finding ways to not need regulated heaters.>

One more round of applause for The HOT Blue Plates (roaring crowd noise)

I admire how Senor Azul always points out the obvious when most of us are pondering what Einstein ate for breakfast?

Salami, salami, bologna   

[shoggoth]

Voltage after the bridge is 6.3x1.414 minus probably 2.2 for the diode drop = approx 6.7.  Then after the 1R resistor, subtract 1R x amps of the heaters after the 1R.  Puts you at 5.8 or whatever.  And again it can all be lower or higher because it's unregulated, and the AC out of your wall may be varying.  6.7v is too low to use a linear regulator to get a 6.3V DC current as HPB notes.  You can use a buck regulator instead.  Switching noise isn't audible, but FCC has regulations around it so you'll want filtering before and after to keep it under control.

[shooter]

Switching noise is something I'm wondering about, in the IC world back when, it was common to put a .1uf cap across B+ n grd about every 10 ic chips.  Would the same work on DC filaments,  say every other tube put a .1 across the filament pins?

thanks

[kagliostro]

Merlin gives a similar council, a small cap in parallel with the heaters pin on V1, this to help to eliminate noise coming from B+, not from the heater PS
K

[shoggoth]

Switching noise is something I'm wondering about, in the IC world back when, it was common to put a .1uf cap across B+ n grd about every 10 ic chips.  Would the same work on DC filaments,  say every other tube put a .1 across the filament pins?

Maybe.  It's a lowpass filter at 1/(2*pi*R*(.1/1000000)), and R is the wire resistance between power source and filament.  The buck regulators I use are 150kHz - the IC's were probably in the MHz.  It's easier to filter MHz noise that way.

An LC filter is a better way to handle it.  The data sheets for the buck regulator IC should have some reasonable examples of a second-order LC filter.

[HotBluePlates]

Switching noise is something I'm wondering about, in the IC world back when, it was common to put a .1uf cap across B+ n grd about every 10 ic chips.  Would the same work on DC filaments,  say every other tube put a .1 across the filament pins?

No, don't waste caps (=money) like that.

Look back at the schematic Ed posted. There are 4700uF caps to filter the heater voltage.

It's a lowpass filter ...

You're not wrong (I know what you're talking about; cap to ground = lowpass), but the intent of their use is to be a highpass filter.

High-value aluminum electrolytic caps generally have good performance passing low frequencies (as they should do in their job filtering 120Hz hum in this case) but they are imperfect devices. They can have a rising impedance past some frequency, such that they do not do a good job filtering 100kHz or so, which could be generated by diode switching.

The 0.1uF cap is to bypass the electrolytic, because it will not have the same issue with rising impedance due to the manner of its construction. This is sometimes done in hi-fi, might be overkill in a guitar amp. Ideally, the cap bypass is placed right at the circuit being powered, while the big electrolytic cap can be remotely located. For a B+ filter cap bypass, this would amount to a film cap from the B+ side of the plate load resistor directly spanning to the ground side of the cathode resistor.

Try using one. If you're brave, take a pair of rubber handled diagonal cutters and snip one leg while someone else is playing through the amp. See if you even hear a difference. (Or if you're smart, maybe put a SPST switch from one leg to where it would conenct in the amp to connect/disconnect)

The basic premise, is if someone is looking at DC filament voltage, one of the things to look at is germanium diodes, and not silicon diodes, you have a couple of  tenths less voltage drop.

Germanium would be lower voltage drop, but in the current ratings required you just can't find them. Schottky rectifiers are available, and would probably be the thing to use. The lowest forward voltage drop I've seen in those (in 1A diodes that seem decent candidates) is 0.55v.

Anyway, the point is that you can use regulation but you have to plan it at the outset of a build; it's not something you can tack onto an existing amp and its 6.3vac winding.

[Ed_Chambley]

HBP Stated:Schottky rectifiers are available, and would probably be the thing to use. The lowest forward voltage drop I've seen in those (in 1A diodes that seem decent candidates) is 0.55v.

Anyway, the point is that you can use regulation but you have to plan it at the outset of a build; it's not something you can tack onto an existing amp and its 6.3vac winding.

This is the reason it was interesting.  A very simple DC circuit that can be added easily and not really need any additional room to do it.


The reason I left the Schottky comment is I knew about the lower drop.  Just forgot to mention it.  I am going to try another using them.  Maybe I will be happy enough with it to try in my one and only high gain amp.  My V-H amp for when I am not using a thumbpick.

[sluckey]

Just a passing thought... Rather than throwing all this dc filament stuff at a Silverface Deluxe Reverb, maybe try to find the real reason that it is noisy? May even be easier.   

[HotBluePlates]

Just a passing thought... Rather than throwing all this dc filament stuff at a Silverface Deluxe Reverb, maybe try to find the real reason that it is noisy? May even be easier.   

That's good advice.

Platefire once used a listening amp (look in the Libary if you don't know what it is) to troubleshoot a hissy amp and found a cathode resistor was making the noise (not the typically-blame plate load resistors).

All my amps have a.c. heaters and don't make noise. The McIntosh MC-30's I have use 2 widely-spaced bare buss wires to connect to the heater pins on the small tubes and have no hum or noise (can't twist them when they're 1-1.5 inches apart).

So most of the time d.c. heat is no needed except in very low-level preamps or microphone preamps where hum has to be zero (d.c. heat usually doesn't fix hiss on its own, without some other fault being present, or it being a secondary effect of the circuitry used to implement d.c. heaters).

[Ed_Chambley]

Just a passing thought... Rather than throwing all this dc filament stuff at a Silverface Deluxe Reverb, maybe try to find the real reason that it is noisy? May even be easier.   

Yes, Sluckey.  I believe you are messing with me.  If not you completely missed my comment about not leaving this in the amp.  Also, I mentioned the amp needs a lot of work.  I just got my hands on it.  It is a 72 and it is all stock.  I will go through it, but I already have a good idea where the problems are.  I got it for cheap, but I have to put it aside for a little while.


Refer to my Reply #6.


I would never leave anything to lessen the value, but I am going to have to rework most of the lead dress.  For some reason the thing earlier type wire.  Gosh darn, I just added a couple of parts.


I promise it is not going to be easy to make this amp look like I will want it.  It is not a repair, it is mine till someone whats it worse which they will when I refurb it.


Again, I just found this interesting and not a substitute for a proper heater string in a Fender.  It is non-conventional.  I have an amp to finish building and no I am not planning to put it into the vox build either.


What is wrong with a little experimentation and discussion?  If it is not cool, just let me know and I will not bring anything non-conventional to the Hoffman Forum.

[sluckey]

Well I missed the point of this thread. Somewhere I got the idea you were trying to fix an old DR. I see I was wrong. Sorry. Please continue.

[shooter]

Most of my latest builds are SS - diodes - n they are quiet, but when I scope power there are a bunch of dancing transients around the 20k range n well they just bother me, nothing to do with sound, some of them big enough to loose scope sink.  I'm working on a bench supply that I want totally clean so when I troubleshoot I'm not adding extra work in the mix.  On the B+ I did add a cap in front of the bridge n it knocked down some of the transients, workin on the filament side now, time, what a great thing to waste!