I am looking for an very reliable solution to bring down my power supply voltage from 34 to 30v. I bought some of these on amazon and one already died and the second is just stuck at 32v. The load is not very high, less than one amp (see the Burning Amp 3 input stage, which I think is about 50-60mA). It needs to be bipolar. While I did buy some bigger regulator kits on eBay with heatsinks and such, why couldn't I use a voltage divider? If I use low enough resistance, like a 1 ohm and an 8 ohm, it should work without affecting the power supply's ability to provide current when needed right? This is for the class A input stage of the BA-3 so it should be pretty constant anyway. The output stage is perfectly happy running at 34v but I want to protect the Toshiba JFets as much as possible.
Most of the amplifiers will support a 10% increase in voltage.
ITOH, you can add a bipolar capacitor in series with the transformer's primary taking advantage on the reactive capacitance, being it non-disipative and no wasting the difference in heat. I used it to convert a Sony radio for 110V into our net of 220V adding a .47uF cap in the primary.
You can also an inductor but they are more big in size, weight and a few of heat is added too.
ITOH, you can add a bipolar capacitor in series with the transformer's primary taking advantage on the reactive capacitance, being it non-disipative and no wasting the difference in heat. I used it to convert a Sony radio for 110V into our net of 220V adding a .47uF cap in the primary.
You can also an inductor but they are more big in size, weight and a few of heat is added too.
I would add a RC stage. This will reduce the voltage and add some filtering as well. Use Ohm's Law to calculate resistance.
LM317 is rated for 37V - very close to the 34V and not a very good idea.
How much current does it draw? If you need to lose 4V, I'd use a string of 6 diodes.
How much current does it draw? If you need to lose 4V, I'd use a string of 6 diodes.
Please state precisely load current.
Is it constant or variable?
Obviously there is something very wrong with your approach,100% mortality is unusual.
"less than 1A" and "50-60mA" are completely different things, please be more precise and state actual value.
Even better, measure it and post here.
And I do not see anything, please post schematic here.
Wrong.
An 8 ohm load (besides your actual needs) draws: 30/8=3.75 AMPERES 😱 on its own, let alone actual needs.
Again, post schematic here.
You are asking, don´t send us googling to answer.
And we might end speaking of different things anyway, so .....
Besides, you don´t have a real problem.
Except a self-created one that is.
I bet the designer designed well to begin with.
But we won´t be certain until we see the schematic.
PS: and that regulator board design is ludicrous.
NO WAY EVER they can supply 1A as they let you imagine, because dissipation is horrible: no heatsinks, SMD package, they will dissipate a little through legs which anyway are NOT soldered to large pads, Output (body) is soldered to a same size pad, where metal gets sandwiched between 2 epoxy layers (PCB and chip case), ugh!!!
I can guess a NO CLUE designer but hey, "anything you make in China sells, so,why bother?"
He should clearly state that it´s a very low current regulator, suitable for preamps and such ONLY ... but I doubt HE even noticed that.
It should be prominently displayed in the Amazon page.
PS: you should have read the reviews: 1 with 3 points, one with 2, stating: "I built a better one" and "lots of ripple in the output"
You should add yours, so others do not fall in the same trap.
Last edited:
Thanks for the replies, let me provide more information as I know it.
I have built the Nelson Pass Burning Amp 3, designed to run at 32v. My front end uses NOS Toshiba 2SK170s and 2SJ74s. Nelson says these are rated at 25v but he wouldn't worry about running them as high as 30v. They release magic smoke around 40v and exhibit much sorter lifespans on the way there. Since the amp is called burning amp for a reason, fearless amp builders can run it at 32v. I paraphrased from the Burning Amp 3 document here. See page 6 specifically. My two Antec 3228 transformers, at my wall voltage, after power supply losses, put out 34.2v. I am using two XRK's Smooth Like Butter power supplies - documented here.
The input section schematic by itself is here: BA 3 Input Stage Schematic
NP says the bias current of the JFETs are about 8mA while the output mosfets run at 45-50mA with peaks of 100mA.
I bought these regulators too: LM317/LM337 Regulator Kit in case those cheap amazon ones failed on me. LOL. Yup, more LM317s but now with heatsinks.
Yes, I did read both reviews, those were bought just to isolate my original problem if the high voltage was the cause (which turned out to be two bad/damaged P channel transistors). I was quite suspect of trying those boards, but they served their purpose and now I need a permanent solution. Or maybe not if 34.2v is fine for those JFETs long term.
While this amp is my 6th DIY audio project, some of the ideas offered here are not immediately understood - I'd be happy to google them and figure it out if something in particular best fit the situation.
My amp is here: Pics (though without the cheap regulators mounted to the back panel)
I have built the Nelson Pass Burning Amp 3, designed to run at 32v. My front end uses NOS Toshiba 2SK170s and 2SJ74s. Nelson says these are rated at 25v but he wouldn't worry about running them as high as 30v. They release magic smoke around 40v and exhibit much sorter lifespans on the way there. Since the amp is called burning amp for a reason, fearless amp builders can run it at 32v. I paraphrased from the Burning Amp 3 document here. See page 6 specifically. My two Antec 3228 transformers, at my wall voltage, after power supply losses, put out 34.2v. I am using two XRK's Smooth Like Butter power supplies - documented here.
The input section schematic by itself is here: BA 3 Input Stage Schematic
NP says the bias current of the JFETs are about 8mA while the output mosfets run at 45-50mA with peaks of 100mA.
I bought these regulators too: LM317/LM337 Regulator Kit in case those cheap amazon ones failed on me. LOL. Yup, more LM317s but now with heatsinks.
Yes, I did read both reviews, those were bought just to isolate my original problem if the high voltage was the cause (which turned out to be two bad/damaged P channel transistors). I was quite suspect of trying those boards, but they served their purpose and now I need a permanent solution. Or maybe not if 34.2v is fine for those JFETs long term.
While this amp is my 6th DIY audio project, some of the ideas offered here are not immediately understood - I'd be happy to google them and figure it out if something in particular best fit the situation.
My amp is here: Pics (though without the cheap regulators mounted to the back panel)
+1
But anyway, barking the wrong bush. The load per channel, correctly working(?), is 53mA.
https://cdn.shopify.com/s/files/1/1006/5046/files/P-BAGSN-1V20-schematic.pdf
Times the 4V drop is 0.43Watt. 0.9W for two. Even the SMD package, soldered to board, can do that. Anyway is supposed to be overheat protected (the real ones are).
The ad does not say what current is supported? (If they don't say, they are not lying!)
"Output short circuit protection. Overheating and overheating protection. Adjust the safety work zone protection."
Failure or wrong output can be bogus parts or mis-application. As kodabmx says, we are close to the LM3x7 peak voltage rating, which applies if there is a load capacitor.
Thanks, NOW we are talking 😉
IF the original +V/-V supply was good enough, I don´t see the NEED to extra regulate it, which simplifies things.
(To boot, I am a Minimalist, so ... 😉 )
1) your Fets can be trusted with an up to 30V supply,so I suggest dropping 5V from current 32Vv rails. Fine with you?
2) then we would need 5000mV/53mA=94 ohm resistors, I´d round it up to 100 ohm each.
Dissipation?: 5V*=294mW, so 1/2 W would be enough, use 1W if you wish.
In series between the +V/-V source feeding this front end at each rail.
I would add some decoupling, say 1000 or 2200uF*35V (or 50) from each "new" rail to ground.
Bonus points:
* dropout can be very low and "always enough"
* absolute worst case: resistors take longer to die, smoke to tell you they are suffering, are cheap and easy to replace, don´t need heatsinks (just separate them 5 or 10mm from board surface)
Thanks JM, I like the simplicity of the solution, it's elegant and reliable. I already got those parts on hand so I can implement it this weekend.