• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

rectifier limiting resistors capacitor input

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi There
Do people routinely use limiting resistors on the rectifier anodes for their capacitor input power supplies?! They seem to be occasionally mentioned as a good idea but I never see them on any schematics!
As a matter of fact I've fitted 2x 100r 12w resistors on my we91 300b p/s (5u4g) as the (Hammond) Tx had pretty negligible resistance. They do run very hot!! I'm wondering about taking them off to save cooking some of the capacitors in there...
Any advice welcome
Thomas
 
I never-ever use resistors in the plates of rectifier tubes. All they do is needlessly waste power as you have found out. As long as your 1st (input) capacitor is 50uFD or less, any self respecting 5U4 should feel no pain. The only time these resistors might help is in the case of a sickly Chicom tube or an OOT Russian tube. But in this case, get yourself a good NOS name brand American or European 5U4.
 
The thing to use is not a resistor but a thermistor. Thermistors are made to limit in-rush current. They are actually temperature dependent resistors. At room temperature they are about 100 or 200 ohms and then after a few seconds of current flowing they drop to 1 or 2 ohms.

Most vacuum tube rectifiers must be used with smaller size caps. The reason for the smaller caps is to limit the total amount of in-rush current. Resistors can limit the current but they are only needed for a few seconds. Thermistors solve the problem. You can use a larger cap if you use a curent limitor of some kind or if might be worth it in some cases.

Limiting the start up current is not a bad idea. It is best to let the tubes heat up before raising the B+ voltage. The thermistor helps here too. They cost about $2.

You can place the thermistor either in series with the primary (limiting the primary curent has the effect of limiting the secondary current.) or you can place them on the secondary or in both places.

You do have to size them carefully. Almost every switching power supply uses current limiting thermistors.
 
Only time I use resistors is making SS replacements for tube rectifiers... calculate the resistor value to equal the voltage drop of the rectifier in question at the specified current draw.

12W resistors will get hot in normal operation. At full dissipation in an enclosed chassis, you can melt solder on them.

Cheers!
 
Administrator
Joined 2004
Paid Member
I try to design my power transformer HV secondaries to provide the minimum required resistance obviating the need for external resistors in most cases.
Adding a UF2007 in series with each plate helps a lot too, whilst keeping the nice soft rectification characteristics of the tube rectifier.

I tend to use a lot more 5AR4 than 5U4 because of significantly the lower forward drop under high current conditions. The 5U4 is marginally more rugged than an old school 5AR4 (Mullard, Tele, Philips) but a lot more so than most modern production I have encountered.

Ignore the minimum winding resistance spec for the 5AR4 at your peril, it may explain why so many of them fail in service, particularly the generally very marginally specified modern production.

Obviously not exceeding the maximum rated input capacitor value helps a lot too, but it is not the entire story. I have seen plenty of instances where tubes lived with 100uF input capacitors and others didn't with just 20uF input capacitors. The difference was winding resistance in many cases.
 
Last edited:
Interesting opinions as ever! Thanks for your replies.
Yes I reckon the two resistors could certainly cook an egg...!
The thermistor idea sounds a promising way to turn down the heat... Which one would you suggest? I did look at the RS webpage (UK), the ones listed as 'inrush current supp' had good current (2 or 3A!) but their initial resistance seemed too low (2-25r) and I wasn't sure what current capabilities the 100r 'ntc thermistors' had...enough for a we91?? Help please! Thanks
Thomas
 
Administrator
Joined 2004
Paid Member
By the way the 5U4G I'm using at the moment is a nos Japanese one- no idea if its good/bad/indifferent... and the input capacitor is a 10uF.
Thomas

I think you are all set, not sure why you think you need the resistor in this application? Have you measured the primary and secondary dcr to determine whether you have a problem?

I am not a fan of NTC on the high voltage secondaries, they aren't really designed for operation at voltages above 250Vrms and if they malfunction at the higher voltage some messy things can happen. (First hand experience)
 
Use PSUDII to find the actual RMS current through the resistors, instead of the average expected rectifier current (it will be much higher than you expect), which will explain the heat they can create. Rectifier data sheets spec'd the minimum plate winding resistances so it was easy for the users to avoid exceeding the maximum peak plate current for a given winding voltage, load, etc. I would avoid exceeding that at all, especially for new production tubes.
 
Hi There
Do people routinely use limiting resistors on the rectifier anodes for their capacitor input power supplies?! They seem to be occasionally mentioned as a good idea but I never see them on any schematics!
Thomas

Hi Thomas ,

Some time ago , I wrote a post about this subject , and how you can do
to calculate those limiting resistors . Perhaps it helps you .
In the rectifier tubes datasheets , you will find a parameter called " minimum
impedance per plate " .

When you know it ,there is a very simple ( but with good results ) formula to calculate the impedance :

Rt = Rs + Rp . N2 where :

Rt = Total impedance ( transformer natural impedance )
Rs = Secondary DC resistance , measured with a multimeter
and :
Rp . N2 = The influence of primary impedance over the secondary
( called reflected impedance )
where :

Rp = Primary DC resistance , measured with a multimeter
N = Transformer ratio ( Secondary Volts / Primary Volts )

I’ll give you an example :

Rs = 35 ohms
Rp = 6 ohms
Transformer ratio ( N ) = 300 VAC output / 117 VAC input = 2.56

Then Rt = 35 + 6 . ( 2.56 )2 = 74 ohms , that way , you’ll need TWO 27 ohms resistors ( preferably , wirewound 5 watts resistors ) between each secondary tap and each plate of the 6BY5GA ( the tube that I had used as an example ), because 74 + 27 = 101 ohms > 100 ohms ( minimum impedance per plate )

If you use this formula with a split secondary ( center tap ) , you must to consider that :
( Rs ) will be the secondary DC resistance between one tap
and the center tap .
( N ) Transformer ratio will be the
HALF total secondary voltage / primary voltage

Did you understand ? You need to complete the transformer
natural impedance with resistors , to reach the minimum impedance value per plate .

Regards ,

Carlos
 
Thanks Carlos thats extremely helpful.
I did actually measure Rs and Rp(to centre tap) when I put the amp together about 4 months ago but can't find the figures now of course, and to be honest, doubt if I used that formula at all- just remember R's being very low. Now I'm paranoid I'm being particularly stupid/worried for no reason and have added resistance where I don't need to... I shall haul the amp down tonight and remeasure...
Incidentally I was thinking of putting the thermistor just past the switch on the primary side- it would see mains (250vac) which seems to be within spec for many of them on the RS website (265vac max?)- just which one?!
Thanks again
Thomas
 
Hi,
the Thermistor location varies with what you want the current limiting to do.

Fit the Thermistor in the primary side of the mains transformer to limit the start up current that the transformer draws at start up from zero core flux.

Fit the Thermistor in the secondary side to limit the charging pulses going through the rectifier to the smoothing capacitors.
 
Ok so fitting one on the primary side won't particularly give the rectifier any protection against initial capacitor current surge. On the other hand, putting them on the rectifier anodes they see 450v they won't particularly like....

Enthusiasm has got the better of me, I've opened up the amp and measured Tx resistances...
Looking at Carlo's formula I've got these figures;
Rp 4.5r
n 400 (ct) /250
Rs 44.5r (per secondary winding to ct)

So I work out Rt to be 56r per anode

Interesingly I get a different answer if I put in figures pretending its not a centre tapped secondary...doh! I thought I was doing so well!

Now I look at the valve data sheets and anode impedence min is 75r or 85r for 450v for the 5u4g. So by rights I should take off my 100r resistors from the rectifier anodes and put in 22r 12w ones instead. Cool. Or running substantially cooler I hope.
Or can I just leave them off and be damned?!
Or, can I leave them off and just cheat and put in say a 100r or so after the rectifier before the capacitor input, so it's an rclclc p/s?!- would that still be 'good practice' for rectifier valve life?!
Comments (not rude ones) welcome!
Thomas
 
Ex-Moderator
Joined 2003
As was mentioned earlier, it is the rectifier peak current that is the problem, and ensuring the correct source resistance keeps it within the rectifier manufacturer's maximum limits. A thermistor won't help. Analyse the circuit in PSUDII and check that your peak current is tolerable according to the rectifier data sheet. If not, increase series resistance.
 
Use PSUDII to find the actual RMS current through the resistors, instead of the average expected rectifier current (it will be much higher than you expect), which will explain the heat they can create. Rectifier data sheets spec'd the minimum plate winding resistances so it was easy for the users to avoid exceeding the maximum peak plate current for a given winding voltage, load, etc. I would avoid exceeding that at all, especially for new production tubes.

For SY's impasse preamp I noted that the power supply resistors were running much hotter than I expected. Total DC draw is about 38mA -- but the ripple measured on the scope was about 2V on the first 470R -- so the I^2 R was a lot higher for AC+DC. The first resistor in the chain ran between 95 and 100C, the second ran at 82C. This was in free air. In even in a ventilated case the temperature would probably rise a lot, perhaps to the failure level. FWIW, I am using fusible resistors these days...

Anyone measure the capacitance of a tube rectifier? -- I get around 10pF for a Sylvania 5U4GB.
 
Disabled Account
Joined 2006
My recommendations:

1. Add an inrush current limiter between the mains switch and transformer primary. I use a CL-190 (Mouser part number 527-CL190). When selecting the device check that A) the max current rating of the device is at least twice the nominal draw of your amp, and B) it's 'on' resistance at your amp's working current will not cause too much volt drop. (Note; you only need one, so ignore those who suggest adding them to both legs of the transformer primary). [ http://gb.mouser.com/Search/Refine.aspx?Keyword=cl-190 ]

2. Add limiting resistors in the anode legs of the rectifiers. Refer to the rectifier manufacturer's datasheet for suitable values. Calculate the values as described by Carlos / refference. Use high-dissipation resistors. Example; if you calculate that you need a 30 Ohm 2 Watt resistor then use 3 x 100 Ohm 2 Watt resistors in parallel for a total of 33 Ohms and 6 Watts. Space the resistors in free-air.


I'm not a tube expert, but doing it this way normally works.
 
Last edited:
I've often seen the PSU11 mentioned and would be really interested to try it out- unfortunately it doesn't work at all well on my vintage imac...drat!
Thanks Gordy for your recommendations which make perfect sense. I will order some cl-190s in a minute. I'm also about to parallel up some 100r's, dropping series R to 33r. I'll see if I lose some of the heat....
Interestingly, the cl-190 thermistor initial resistance being 25r, would according to Carlos' formula, render series resistance sufficient without the use of anode resistors, as series resistance becomes around 120r per anode, well past the 75 or 85r minimum according to data sheets...(am I interepreting that right?!)- at least for the split second the thermistor doesn't conduct...
I think ideally I would like to eliminate the anode limiting resistors as they are only needed on start up and burn up power. With them gone perhaps I could reduce the input capacitor a little....
Thanks again
Thomas
 
Just about the same for a cold 12CL3 in a socket.

I never paid any mind to it, nor was it ever mentioned long those many decades ago, but we all had this type rectifier or some near-cousin in our Halliscratchers, Hammarlund, Heath and Eico stuff on the bands from 160 to 10M -- the RLC circuit formed by the rectifier tube, the leakage inductance of the trafo and the associated resistances all had some pernicious stuff going on and may have been associated with some of the fuzz one heard on the ham and SWL bands.

I put a 5U4GB warmed up on my HP3577 spectrum analyzer with a purely resistive load approximating what one would encounter in a ham receiver of the era -- and there is a point where the phase goes wonky just like a tuned circuit.

I dunno, probably doesn't account for a hill of beans...
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.