scott2000
Well-known member
Maybe a diagram of how you've hooked up the transformer would help too.
DC Filament Supply questions
- Thread starter jonasnoble
- Start date
Help Support GroupDIY Audio Forum:
I don’t think you need more than 12.6V at 2A at all - I was suggesting to check all voltages through every stage with a dummy load in place - but the low DC voltages you’ve been tabling indicate that there is a flaw somewhere in the layout(s).
78xx series regs run hotter as the voltage climbs so it’s best to keep the voltage at around 3V above required minimum input as a max - that way the only thing the regulator will boil is water. If the ripple in the input voltage from the bridge swings above and below the minimum voltage required to turn the regulator on you will get unstable performance and lack of current delivery. Pre-stabilising with a DC-DC converter to cover the drop from whatever the bridge delivers with the secondaries in series and feeding say at least 9V into a 6V reg would keep the stress on the components to a minimum and allow a margin in the event there is a fluctuation in mains voltage.
What is the usual tolerance on heater filaments? I always see 6.3V nominal, so I was going to suggest just using AC filament supply, but I see that the transformer is rated for 6.3V secondary at 115V primary. Where I live it is common to have 124V line voltage, at least in at certain times of day. That would result in around 6.8V secondary voltage, about 8% high.
Traditionally it was good practice to design for same performance at +/-10%, and acceptable performance at +/-15% or 20%, so I would hope tubes can take +10% without any problem. I don't have any tube datasheets available to see if that is specified though.
But given the secondary voltage that transformer does look like it was designed with AC filament supply. Is this entire exercise just overcomplicating something needlessly?
Traditionally it was good practice to design for same performance at +/-10%, and acceptable performance at +/-15% or 20%, so I would hope tubes can take +10% without any problem. I don't have any tube datasheets available to see if that is specified though.
But given the secondary voltage that transformer does look like it was designed with AC filament supply. Is this entire exercise just overcomplicating something needlessly?
Something needed to help
How many and what tube(s) will be used? This will help with knowing the current needed.
This can matter a lot depending on the current and DC noise you can work with.
This thread seems to indicate you will need to run the output at 12.6VAC into a diode bridge OR at 6.3VAC into a voltage doubler depending on the current required
All you might need after the bridge is a cap, series dropping resistor(calculate resistance and power at worst case for >3VDC before the regulator), cap, regulator and cap. You will need a heatsink on the regulator
How many and what tube(s) will be used? This will help with knowing the current needed.
This can matter a lot depending on the current and DC noise you can work with.
This thread seems to indicate you will need to run the output at 12.6VAC into a diode bridge OR at 6.3VAC into a voltage doubler depending on the current required
All you might need after the bridge is a cap, series dropping resistor(calculate resistance and power at worst case for >3VDC before the regulator), cap, regulator and cap. You will need a heatsink on the regulator
Last edited:
Thank you!
Currently one tube, 6N1P draws 600ma on the heater. I may add another 12a_7 at some point, but should still come under 1A.
I will simplify the circuit and start scoping tonight. Any recommendations on sizing the filter caps? Maybe I don't have high enough values, but sims didn't see any impact by changing values.
Before thinking too far ahead, or bothering with many sims or whatever, just start by measuring what that transformer of yours is ACTUALLY putting out (or not). Each secondary on its own, and then both secondaries in parallel. Then connect the bridge rectifier, and verify you get the same AC RMS reading.
At this point, you could consider connecting a dummy load that sinks ~600mA for the voltage you measured previously - mind your Ohm's law, and remember to also calculate (at least roughly) what sort of power the dummy load will need to burn off, and size your resistors accordingly. Series-parallel combinations, if needed.
Then connect whatever smoothing capacitor you wish / have (1000uF is a start, might wanna go a bit higher, but anything over 4700uF gets borderline redundant, especially if you'll be using a regulator after that), and verify that you get more or less the AC voltage times 1.4 (give or take). Now take DC voltage readings with and without the dummy load. Optionally you can also now scope (AC-coupled) the voltage across the smoothing capacitor, with and without the load, and see what amount of ripple voltage there is - probably won't be much without the dummy load.
I would suggest putting an AC current meter in series with your tube filament and see what it draws running on AC from that transformer at 6.3V also measure the AC voltage across the filament. This will allow you to see what value dummy load you need for testing the DC circuit. (As I mentioned before a 10W wire wound should be more than sufficient or even a 5W for the same load as the filament should be but it might get pretty hot 
.)
There is always a possibility that one of the windings is faulty - as Khron mentioned in #38 the two windings should be measured loaded for their AC voltage output. It might be a bit tricky getting a single resistor that will have a high enough wattage to test at 2A but a pair of 6.8Ω 10W WW should be easy enough to find (8Ω also easy), put in parallel will draw 1.85A and used singly will give a higher load than the tube should be drawing - you may want to leave a margin below the 2A so you don’t blow the transformer. My local electronics store sells most preferred values of wirewound resistors so I’ve never needed to wait for online supply.
You don’t want to go into the DC side until you’re sure of your transformer.
.)There is always a possibility that one of the windings is faulty - as Khron mentioned in #38 the two windings should be measured loaded for their AC voltage output. It might be a bit tricky getting a single resistor that will have a high enough wattage to test at 2A but a pair of 6.8Ω 10W WW should be easy enough to find (8Ω also easy), put in parallel will draw 1.85A and used singly will give a higher load than the tube should be drawing - you may want to leave a margin below the 2A so you don’t blow the transformer. My local electronics store sells most preferred values of wirewound resistors so I’ve never needed to wait for online supply.
You don’t want to go into the DC side until you’re sure of your transformer.
Shame they don’t do a 6V one - 5V or 9V are the choices down there. Looks like he could use the LMS78_9-2 to 9V then the 7806 linear to drop to 6V and isolate any switching noise with filter caps between the two as well as on the bridge rectifier and the 7806 output. I’d be careful of any switching noise from the LMS78 radiating from the filament and it’s wiring to the rest of the circuit - it has from 30mV to 75mV ripple and noise. They have an output capability of 2A so well suited for this (they cannot be used in parallel to up the output current). The 7806 needs to be spaced away so the temp doesn’t go above 85deg or the output power drops like a stone.
They table the recommended capacitor values for input and output in the datasheet - 22uF/50V in and 22uF/16V out - ceramics.
How about something like these for a compact but plenty-capable-enough solution to drop just about anything down to 9V?
https://www.aliexpress.com/item/4000975258784.html
It would require replacing one of the two feedback resistors though...
https://www.aliexpress.com/item/4000975258784.html
It would require replacing one of the two feedback resistors though...
Or these - they’re adjustable for voltage and current limiting:
https://core-electronics.com.au/dc-...Jnk8HAc0PHmMQp2soBZavei1M0K8kipkaAhSLEALw_wcBLess than $5!
https://core-electronics.com.au/dc-...Jnk8HAc0PHmMQp2soBZavei1M0K8kipkaAhSLEALw_wcBLess than $5!
That's so expensive!
The one on Aliexpress that i linked is 4.41eu for a 10-pack (and another 1.41e shipping...)
Galdimmidy, dang who’d a thunk - I’d best saddle up the pony and get mahself a bundleThat's so expensive!
I’ll have to calculate the time to modify them to get the right voltage and do a costing spreadsheet, adding the time for creating the spreadsheet and see which is more cost effective 
The chip they're based on seems pretty capable for its size (8-30V in, up to 2.8A continuous output)...
http://www.sinpie.cn/uploadfile/20160902/20160902213919235.pdf
Probably won't wanna go above 15V out due to that black tantalum(?) output cap though, but still, for the price......
http://www.sinpie.cn/uploadfile/20160902/20160902213919235.pdf
Probably won't wanna go above 15V out due to that black tantalum(?) output cap though, but still, for the price......
Guys this is fantastic info. I spent my commute this morning learning how buck converters work. Just bought these off amazon to experiment.
https://www.amazon.com/gp/product/B079N9BFZC/ref=ppx_yo_dt_b_asin_title_o01_s00?ie=UTF8&psc=1
$15 US for a 4-pack.
I've never ordered from AliExpress before, but when I went to order a new spindle for my CNC router yesterday, they did some slight of hand, and I almost paid for something I didn't want or need. I don't think I trust the prices until I click the buy button.
https://www.amazon.com/gp/product/B079N9BFZC/ref=ppx_yo_dt_b_asin_title_o01_s00?ie=UTF8&psc=1
$15 US for a 4-pack.
I've never ordered from AliExpress before, but when I went to order a new spindle for my CNC router yesterday, they did some slight of hand, and I almost paid for something I didn't want or need. I don't think I trust the prices until I click the buy button.
I think they’re a brilliant solution and with the right filtering you could use them on anything as long as you put a scope on the rails to see if there’s any noise.
With AliExpress you need to read carefully - when you go to the choose page you select your “color” which means which part of a bundle you might be buying - you might see a drill press for $2 but that turns out to be the little clamp that holds it to the bench, the press is $30! Once you get used to it, it’s fine.pack.
I've never ordered from AliExpress before, but when I went to order a new spindle for my CNC router yesterday, they did some slight of hand, and I almost paid for something I didn't want or need. I don't think I trust the prices until I click the buy button.
I once bought what I thought was a VCO chip, originally sold through Radio Shack, price was good - until I received the package which contained an original Radio Shack datasheet, faded and yellowed but in pristine condition - searched the package - no IC, just a datasheet. Hmmmm
haha, that's almost the same as what happened to me. I was shopping for spindle motors and saw a $10 ad on Google. I figured it was some introductory offer to get me to sign up. I created an AliExpress account and went to buy the spindle kit, and right before I hit the "complete purchase" button, I noticed they had switched it out for just the bracket.
