Elevated heaters in new build, placement of Standby switch

[PharmRock]

I've been checking out various forums and tech info sites (e.g. Paul Ruby) and reading about the benefits of elevated heaters and how it may decrease hum, especially in a higher gain amp.  Since I basically have a "clean slate" to go with, I am seriously considering incorporating this into my 50W Plexi/800. 

I found a couple of great examples on this thread here:
https://el34world.com/Forum/index.php?topic=15932.msg154820#msg154820,
Tubeswell has an example in the above thread where he basically set up the heater elevation with the bleeder resistor network.  I think this design is a great way to accomplish the heater elevation and a bleeder with minimal parts/space.  I have a few questions about implementing this in my build:

1. I am using a ClassicTone PT that puts out ~690VAC on the primaries and has a center tap for heaters.  I'm figuring I should hit ~450VDC after diode rectification.  If I use a voltage divider of 470K/56K at the first filter node (50uf+50uf), it should put me around 50VDC for the heater elevation.  Based on some research I found on this forum https://el34world.com/Forum/index.php?topic=23285.0 it appears a 1W rating for both of the resistors should suffice.  Am I calculating the resistor values and power requirements correctly?

2. What value capacitor should I use for the elevated heater supply voltage voltage divider?  I would think I would need a voltage rating equal to 2x the voltage (e.g. 100V), but what about the capacitance value?  I've seen anywhere from 16-100uf in examples I've found.

3. My amp will have a standby switch.  Based on the "regular" 50W plexi schematics I've read, the standby switch would normally be placed between the diodes and the first filter node (50+50).  But it seems to me if I do this, while in standby the heater CT would not be referenced to the 50VDC voltage since there is no DC at the first cap yet.  Would the heater CT connection to ground via the 56K voltage divider resistor be ok while in standby (and allow the heaters to run at 6.3VAC while in standby), and then when I switch from standby to "on" position, the heaters would be then be elevated at 50VDC? 

thanks for any comments/suggestions.

[pdf64]

Elevating the heater circuit does not affect the heater VAC.

[PharmRock]

so if I have the elevated DC supply at the first cap and before the standby, when I'm in standby mode I have a "normal" heater situation but with the CT connected to ground via my 56k voltage divider resistor (vs directly connected to the ground bolt in a typical set up). 
When I switch from standby to "on", I am now sending DC voltage to the first filter cap, and at that point my heaters will be then be elevated at 50V via the 460k/56k voltage divider. 

otherwise no problems with my approach/calculations?

thanks for the help

[2deaf]

I would rather have the heaters elevated all the time.

[d95err]

Leave the elevation on all the time.

Also, you want the voltage for the elevation to be as clean as possible. The first filter stage can have something like a 5V ripple.

I use the voltage from the second or third filter node.

[pdf64]

Standby is a nonsense, the endless faffing around in order to accommodate the constraints and complications it introduces seems topsy turvey to me 
Just get rid of it.
If there’s a panel hole for the switch, use it for something else.
We’re (hopefully) happy to do that for the anachronism that is the ground switch, so I’m at a loss to understand the reluctance to do the same for standby 

[PharmRock]

I would rather have the heaters elevated all the time.

Leave the elevation on all the time.

This makes sense to me, and was my main concern with having the voltage divider running off the bleeder resistor network AFTER the standby.  If I did it that way, I would be running the heaters referenced to 0V in standby, and at ~50V during operation.  Didn't seem right.

Standby is a nonsense, the endless faffing around in order to accommodate the constraints and complications it introduces seems topsy turvey to me 
Just get rid of it.
If there’s a panel hole for the switch, use it for something else.
We’re (hopefully) happy to do that for the anachronism that is the ground switch, so I’m at a loss to understand the reluctance to do the same for standby 

With respect to the standby/no standby switch...I've done a lot of reading on it and it still seems to me there are 2 "camps" on the issue.  The technical aspects exceed my understanding of the subject however.  Many of Doug's layouts incorporate a standby switch, for example, the Plexi 50 layout that I am generally following.  I will be using the progressive switch Doug sells for power/standby, so there is only one hole.  And while this is a minor point, I've already had my front panels made and "standby" is already on there.     With my OCD, it would drive me crazy having "standby" on the front panel when in fact there is no standby option.  Logical?..no.  But that's the reality. 

Also, you want the voltage for the elevation to be as clean as possible. The first filter stage can have something like a 5V ripple.

I use the voltage from the second or third filter node.

This sort of brings me back around to the standby switch issue.  It seems this wouldn't be feasible if I am incorporating a standby switch into the design of the amp.  Otherwise, I'm back at 0V reference at standby, 50V reference when on. 

With 2deaf's schematic, tapping the HV secondary prior to the rectifier diodes and before the standby switch, and using a 1N4007 diode to do the rectification solely for the elevated heater reference, this may be the best option.  But getting to d95's comment about 5V ripple...will this not also occur with 2deaf's design?

Finally, are there any fuse protection considerations with this?  My HT fuse will be after diode rectification, and thus the "tap" off of the secondary for the elevated heater reference will not be "protected", except for the mains fuse. 

I appreciate everyone's help and responses...I'm still very much a struggling student in all of this, but wanting to learn as much as I can.
 

[shooter]

With my OCD, it would drive me crazy having "standby" on the front panel when in fact there is no standby option.

put the switch in, no wires, 
I don't even play but i find lots of things to occupy me while tubes warm up, adjust strap, check switches, fiddle with pick, strum a couple times, then..............jam!!

[2deaf]

With 2deaf's schematic, tapping the HV secondary prior to the rectifier diodes and before the standby switch, and using a 1N4007 diode to do the rectification solely for the elevated heater reference, this may be the best option.  But getting to d95's comment about 5V ripple...will this not also occur with 2deaf's design?

d95err's 5V ripple gets reduced to 532mV by your voltage divider and then it is smoothed by the 16-100uF capacitor to a totally insignificant ripple.

2deaf's 450V ripple gets reduced to 50V and then it is smoothed by a 47uF capacitor.  Tau is hard to calculate, but it is obviously large resulting in maybe a 100mV ripple.  The heaters already have a 3.15V ripple and +/- 100mV doesn't scare me.

Finally, are there any fuse protection considerations with this?  My HT fuse will be after diode rectification, and thus the "tap" off of the secondary for the elevated heater reference will not be "protected", except for the mains fuse.

The bias circuit doesn't have any fuse protection, either.  They both have a large resistor between the secondary and anything that might go wrong, so I'm not scared.     

[PharmRock]

This definitely sounds like the way to go.

The Plexi board I have has two 100ohm resistors in place for an artificial center tap for the heaters.  Since my PT has a heater center tap, I'm going to remove the artificial CT resistors from the board, add a couple of turrets, and have the heater elevation circuit on the board.  This will be convenient because the rectifier diodes are immediately adjacent to this, and I can tap right off of the HV secondary as it enters the board.  Everything will be nice and tidy, no extra terminal strips or off-board components necessary, and components will be in a place already dedicated to the heater center tap anyways.

Working on a layout to make sure I have all this correct. 

Thanks for all the help.

[PharmRock]

I've attached a photo of the board as it is currently, and a layout of the changes mentioned in my previous post.  I only included the relevant portions in the layout vs. recreating the entire bias supply and power amp ground buss.
 I think this will end up working out pretty well.

[PharmRock]

About to install all this on the board, and I just have a question about voltage dividers and diodes, and the power supply in general.

The transformer supplies 690VAC at 150mA. I am tapping the elevated heater supply off of one of the PT secondaries which is 345VAC.

I am using 2deaf's recommendation (see above), and was hoping someone could walk me through the math, like in "Remedial tube amps 101" language.

345VAC hits a 220K resistor (I am using a 2W resistor).  There is voltage drop across this resistor...what would it be on the "other side"?  Then this lower voltage goes through a 1N4007 diode, which will rectify the AC voltage to DC.  I'm trying to figure out what this DC voltage will be at this point.  I've read that these diodes drop 0.7V....is that total or is it in proportion to whatever is supplying it?  On the "other side" of the diode I will have an 82K resistor (all I have is 1W...hope this is ok).  In parallel with this 82K resistor is a 100uf/100V cap.  The heater CT will connect at the junction of the diode and cap+82K resistor. 

I'm just trying to have an elementary knowledge of what is happening once I tap the 345VAC secondary for this elevated heater supply.  Thanks for any replies/comments and for "dumbing it down" to a non-engineer. And thanks 2deaf for supplying the schematic.

[sluckey]

1... Convert 345V_RMS to peak voltage. So, 345 x 1.414 = 488V_PK. (V_PK = 1.414 x V_RMS)

2... Ignore the diode voltage drop since it's forward voltage drop will ALWAYS be .7V.

3... That leaves two series resistors of 220K and 82K. Total resistance is 302K.  (R₁ + R₂ = R_T)

4... I = E/R = 488/302,000 = 1.6mA  (Since the resistors are in series, the same 1.6mA current flows thru each resistor.)


5... 1.6mA current flowing thru 82K will drop 131V across the 82K.  (E = IR)

The cap across the 82K resistor will charge up to the peak voltage across the resistor and that will be 131V_DC. If you want less voltage, decrease the size of the 82K or increase the size of the 220K. Then repeat steps 3, 4, and 5 of the above procedure.

[PharmRock]

thanks for the explanation....I have one remaining question, and that is, where is the AC to DC conversion and resulting "voltage loss" from this conversion?  I know the diode is converting the AC voltage to DC, and I've read that the DC output will be around a theoretical 40% of the incoming AC voltage.  So wouldn't the 131V in step 5 actually be 40% of that....or ~52VDC?  This value seems more in line with elevated heater supply voltage.  My cap is rated for 100V, so based on the 131V I will need to either lower the 82K to perhaps 47K or so, or get a new cap.

Again...just trying to learn along the way, and I appreciate you taking time to "teach".

[sluckey]

Haha! Disregard everything I said above. It's wrong. Maybe one of the smart guys will straighten me out.

So, sometimes it's just easier/faster to go to the bench to sort this stuff. This is with real components cobbled together and actual voltages measured with a Fluke meter. I used a Variac and PT to set the input to exactly 345VAC and measured 55VDC across the cap.

[acheld]

Is the difference due in part to the voltage in

Convert 345VRMS to peak voltage. So, 345 x 1.414 = 488VPK. (VPK = 1.414 x VRMS)

is peak to peak, which then gets cut in half by the diode?

Asking, not stating . . .

[PharmRock]

So, sometimes it's just easier/faster to go to the bench to sort this stuff. This is with real components cobbled together and actual voltages measured with a Fluke meter. I used a Variac and PT to set the input to exactly 345VAC and measured 55VDC across the cap.

This is why this forum is hands-down the best tube-amp forum on the internet.  thanks for the response and "real life" verification. This is awesome.

[sluckey]

Is the difference due in part to the voltage in

Convert 345VRMS to peak voltage. So, 345 x 1.414 = 488VPK. (VPK = 1.414 x VRMS)

is peak to peak, which then gets cut in half by the diode?

I only converted to peak, not peak to peak.

[Sonny ReVerb]

I think this is the math:
V_DC = 0.45 x V_RMS
0.45 x 345 = 155 V_DC
155 / 302000 = 0.5 mA
0.0005 x 82000 = 42 V_DC
Not sure about the actual test results. Could the diode placement between the resistors be a factor? Did the resistors measure accurately?

[PRR]

Haha! Disregard everything I said above. It's wrong. ....

This type supply can NOT be computed by mere mortals. Or rather: it is FAR easier/surer to plagiarize a plan and tinker to needs than to work out what the fudge is really going on in there.

This is NOT a "peak-catching" rectifier. It averages. But much of the cycle it is back-biased and loafing. How much of a cycle? And it can take many-many cycles, even many seconds, to come near a final value. We "can" figure it, but it is not a good use of limited brain-power.

[PRR]

PSUD will compute this contraption with fair accuracy (generally far better than ad hoc theory, though not as good as tacking one together to test). Just remember that PSUD only computes "positive" and you have to read and build opposite polarity. (Actually since PSUD is "groundless", polarity is meaningless; we like to think the bottom line is a common wire but that's a careless assumption.)

[PharmRock]

Thanks PRR...this is pretty cool.  Seems like everything is leading to the same results. 

As far as what you used in PSUD....does the diode placement matter? (i.e. 2deaf had the diode AFTER the 220k resistor in his schematic).

Thanks!...I'm starting to feel more smarter just by looking at this stuff. 

[PRR]

> 2deaf had the diode AFTER the 220k resistor

Your car needs a wash and a fill-up. Does it matter which order?

(OK, sometimes it does. If car is really dry, better gas-up first. If fuel cap is so dirty you'll get dirt down the tank, wash first. And here, if I gas-up I get a chit for $1 off my car-wash. If there is a side-chore between gas and wash, or a sideload between R and D, then order matters. But in general parts in series with no side-path off the junction can go in either order.)

[2deaf]

I think this is the math:
V_DC = 0.45 x V_RMS
0.45 x 345 = 155 V_DC
155 / 302000 = 0.5 mA
0.0005 x 82000 = 42 V_DC
Not sure about the actual test results. Could the diode placement between the resistors be a factor? Did the resistors measure accurately?

So then you're just multiplying the rms voltage by 0.45 then multiplying that by the voltage divider.  I'm going to call that 0.45 factor "F".  F ranges from 0.5 to 0.6 for the voltages that we are likely to desire.  As the load resistor (82K in this case) increases, F increases.  F is right around 0.54 with a 220K dropping resistor (R_d) and an 82K load resistor (R_l).

V_dc = V_rms  x  F  x  R_l / (R_l + R_d)

V_dc = 345V x 0.54 x 82K / (82K + 220K) = 51V

If R_l were 165K, then F would be around 0.6 and I would expect around 89V_dc.

 

[Sonny ReVerb]

The formula in my previous post is a link to a page that shows the math. At least it gives you a starting point.

... in general parts in series with no side-path off the junction can go in either order.)

That's what I thought, but limited brain power and all...

[PRR]

The formula in my previous post is a link to a page that shows the math.

That page does not cover this DC supply. Most of it has no cap, is making a lump-wave. The last plan has a cap amd makes sorta-smooth DC but there is no series resistor. In general this type supply only appears in EE quizes, to torture the students.

[Sonny ReVerb]

Sorry PharmRock, that demonstrates why I didn't make it through EE school. I'll go back to asking questions...

[PharmRock]

Sorry PharmRock, that demonstrates why I didn't make it through EE school. I'll go back to asking questions...

hey I can guarantee your knowledge in all this is orders of magnitude higher than mine! But unless I'm misunderstanding, the equation VDC = 0.45VRMS is pretty close to the "F" factor 2deaf described.  Maybe I am over-simplifying it, but it seems to get in the neighborhood.  The actual value depends on other components such as dropping resistor value, supply voltage, etc.

But much of the cycle it is back-biased and loafing.

loafing...finally a term I can understand.

What I've found most interesting about this is that the tapping off of one of the secondary leads for this elevated supply isn't a whole lot different than what we are doing for the bias supply, except maybe for the "orientation" of the diode and caps and a variable resistor.

Its been a busy past few days, but this will be built out on my board tonight.  Hope to make a lot of progress on the amp this weekend.  Once it gets powered up the first time I will report back with the actual voltages.  Appreciate all the help!

[Mike_J]

Have attached a picture of the elevated heater supply circuit for my 5f6a build. It has a 47uF @ 100VDC radial ecap. R1 and R2 are 430K and 100K respectively. Used 100K for R2 because I was told less mA were used (may be misremembering again). Voltage is taken from the same supply point the screen uses for added filtering. I isolated the 100 ohm resistors on a vertical terminal strip which connects to the positive side of ecap mentioned above. Have done it that way so the amp will only be grounded to chassis in one spot at the input jack. Actually near the input jacks which were lifted to avoid multiple ground points.

[d95err]

Regarding the original question - I remembered that I have one build that has heater elevation and a standby switch (the amp I use the most, actually). The standby switch will disconnect the heater elevation voltage in standby mode. Never even thought about it.

It is no problem at all. No pops or other artefact when switching. No problems with the amp after a few years with rather frequent use.

[PharmRock]

Regarding the original question - I remembered that I have one build that has heater elevation and a standby switch (the amp I use the most, actually). The standby switch will disconnect the heater elevation voltage in standby mode. Never even thought about it.

It is no problem at all. No pops or other artefact when switching. No problems with the amp after a few years with rather frequent use.

thanks for the info...I ended up going with a tap off of one of the primaries, prior to the standby so the heaters are always elevated.  Did you do yours as part of a bleeder network off of the first filter cap?

[d95err]

Regarding the original question - I remembered that I have one build that has heater elevation and a standby switch (the amp I use the most, actually). The standby switch will disconnect the heater elevation voltage in standby mode. Never even thought about it.

It is no problem at all. No pops or other artefact when switching. No problems with the amp after a few years with rather frequent use.

thanks for the info...I ended up going with a tap off of one of the primaries, prior to the standby so the heaters are always elevated.  Did you do yours as part of a bleeder network off of the first filter cap?

I hope you mean tap from one of the secondaries... 😀

I think I tapped off the PI node.  From what I recall, that’s what Kevin O’Connor recommends in one of his books.

[PharmRock]

I hope you mean tap from one of the secondaries... 😀

I think I tapped off the PI node.  From what I recall, that’s what Kevin O’Connor recommends in one of his books.

Well that's a pretty significant typo...yes...secondaries.     I've been staring at this stuff too long!  I have it correct on my layout.