EMRFD Message Archive 4535
Message Date From Subject 4535 2010-04-20 14:24:55 Vojtěch Bubník problem with IRF510 push-pull PA
I built the Altoids PA kit from Makoto Kasahara JN1GLB. It is a multi band design with pluggable band modules.
http://ok1hra.nagano.cz/2008_0228_altoids_pa_assembly_manual_e.pdf
I am experiencing good efficiency at 20m, but poor efficiency at 40m and near zero amplification at 80m. It seems to me that the output transformer has too low inductance for lower bands or it saturates or both. I know the rules of thumb: inductance of the transformer 5x higher than load impedance and if the transformer saturates, it is no more linear and produces heat.
Is there any coincidence with low impedance of the transformer and core saturation? Is it possible to have too low transformer impedance and not to saturate the core? What happens if the output transformer has too low impedance? Can I diagnose it with scope?
I am also seeing some oscillations with pretty high amplitude at both gate and drain of the IRF510 MOSFETs. I was prepared to see deformed sinusoid on gate, but these are clear oscillations. Is it healthy?
Thanks, Vojtech AB2ZA4537 2010-04-20 15:10:17 ajparent1 Re: problem with IRF510 push-pull PA 4543 2010-04-21 08:43:43 ajparent1 Re: problem with IRF510 push-pull PA
Based on several other similar IRF510 amps the bias setting method is
weak as described in the manual. I've built several based on K5OOR,
WA2EBY and KD1JV all being close in design. Their manuals may be helpful.
I'd opt for forcing tx mode using jumper and NO RF suppled and the inputs and outputs terminated with a 50 ohm dummy load. A current meter with (100ma) is inserted in series with the DC line (insert
is series with L2) and set the bias resistors for 10mA per IRF510.
since both are connected the easy way to do this is set both resistors for minimum and then advance one for 10mA then advance the other for a total of 20ma.
It should be stable with no drive, any sign of oscillation needs
attention as to cause. A good test is to drive with about 200mW
as that will not attain full power but should excite it to about
2-5W output with a clean signal. Any instability is suspect.
I now have two IRF510 designs one based on WA2EBY and that runs
more power (typical 40+W) as it was design for 24-28V operation
and the KD1JV (very close the GLB design) versi4549 2010-04-22 19:20:09 Vojtech Re: problem with IRF510 push-pull PA
Thanks for your comments.
I learned that the kit I built is basically KD1JV amplifier modified slightly by K0LR. KD1JV recommends bias of 10mA per FET. I adjusted the bias for current of 100mA per FET. This is very much the same value K0LR recommends. At higher bias current the MOSFET voltage to resistance curve is steeper and the amplifier provides more amplification.
I don't see any signs of oscillation without signal. I made screenshots of voltages at gate and drain. I see some oscillations, but I am not sure how much they are caused by Russian scope probes of unknown quality or whether they are real.
http://www.kufr.cz/~ok1iak/temp/scope-drain.jpg
http://www.kufr.cz/~ok1iak/temp/scope-gate.jpg
Yesterday I measured DC power and RF power generated at various bands. At 20m I am getting 18W with 35W DC current. At 40m I am getting 18W with 45W DC current. At 80m I am getting 9W with 45W DC current. I suppose it is a good indicator that the output transformer does a bad job, most probably saturates slightly at 40m and heavily at 80m.
I have populated boards for HF Packer amp. I suppose I shall bolt it on a heat sink and give it a try. I can compare the two amps then and see what will be different.
Thanks and 73, Vojtech AB2ZA4550 2010-04-23 01:28:22 Weddig, Henning-C... Re: problem with IRF510 push-pull PA
this reminds me to a "rule of the thumb" that the inductive resistance
of the transformer at the lowest frequeny of use should be at least for
times the lowest "real" impedance to be transformed. So as the primary
of the transformer is a very low impedance either the µ of the binocular
is not high enough therefore the number of turns for the primary (and of
course the secondary turns ratio) have to be more than just one turn!
Vojtech schrieb:
>
>
> Hello Allison.
>
> Thanks for your comments.
>
> I learned that the kit I built is basically KD1JV amplifier modified
> slightly by K0LR. KD1JV recommends bias of 10mA per FET. I adjusted the
> bias for current of 100mA per FET. This is very much the same value K0LR
> recommends. At higher bias current the MOSFET voltage to resistance
> curve is steeper and the amplifier provides more amplification.
>
> I don't see any signs of oscillation without signal. I made screenshots
> of voltages at gate and drain. I see some oscillations, but I am not
> sure how much they are caused by Russian scope probes of unknown quality
> or whether they are real.
>
> http://www.kufr.cz/~ok1iak/temp/scope-drain.jpg
> <http://www.kufr.cz/~ok1iak/temp/scope-drain.jpg>
> http://www.kufr.cz/~ok1iak/temp/scope-gate.jpg
> <http://www.kufr.cz/~ok1iak/temp/scope-gate.jpg>
>
> Yesterday I measured DC power and RF power generated at various bands.
> At 20m I am getting 18W with 35W DC current. At 40m I am getting 18W
> with 45W DC current. At 80m I am getting 9W with 45W DC current. I
> suppose it is a good indicator that the output transformer does a bad
> job, most probably saturates slightly at 40m and heavily at 80m.
>
> I have populated boards for HF Packer amp. I suppose I shall bolt it on
> a heat sink and give it a try. I can compare the two amps then and see
> what will be different.
>
> Thanks and 73, Vojtech AB2ZA
>
>4551 2010-04-23 06:29:31 ajparent1 Re: problem with IRF510 push-pull PA 4553 2010-04-23 07:15:00 ajparent1 Re: problem with IRF510 push-pull PA 4570 2010-04-25 13:32:43 david cripe Re: problem with IRF510 push-pull PA
73 Dave NM0S4572 2010-04-25 15:09:15 ajparent1 Re: problem with IRF510 push-pull PA 4573 2010-04-25 16:38:22 bobtbobbo Re: problem with IRF510 push-pull PA
http://openhpsdr.org/wiki/index.php?title=PENNYWHISTLE
Bob, K1AO
>
> > There are a lot of issues with this circuit that could result in oscillation. The secondary of the drive transformer should have a grounded center tap, otherwise you can get a common mode oscillation. Be careful with the lead length between the output transformer and the first shunt cap in the LPF. Ideally, this cap should be directly between the drain and source of each IRF510, with zero lead length. You may need to add a small value resistor in series with each IRF510 gate - 2.2 or 3.3 ohm.
> >
> > 73 Dave NM0S
> >
>
> Actually Wa2EBY used a unbalanced to balanced phase reverser(Un-UN)
> with better result as it's easier to predict a transmissionline transformer behavior.
>
> The real problem is the Cg (gate input capaitance) is spec'ed at 180pF
> and it's DC resistance is infinite so your translating a largely capacitive load to a matched impedance.
>
> Putting some resistance there will damp any oscillation but the gate at 20M look like a 6 ohm reactive load so end up heating the resistor but not driving the gate. at 10M it's down to 3ohms. It's better to load the gate resistively as a parallel load at some value in the 30-100ohm range and then absorb the input capacitance as part of the matching network or work at maintaining lower drive impedance as that will make the drain to gate capacitance less of a factor and aid stability.
>
> The output network (transformer) is likely short on reactance
> at low frequencies. Likely works better as a result at 20-10M
> but this is always a problem for wide band amplifiers at high
> power. More ferrite and some drain to drain capacitance to tune
> out leakage reactance would be good.
>
> Both the input and output conditions are worse when trying to
> run at 12V than 24-28V where the device behaves far better.
> This is reasonable as the device is a 100V power FET.
>
> In the files area under my call(KB1GMX) is the Datasheet, Apnote and WA2EBY's two part article detailing this and a successful amplifier.
> All attempts to build a wide band amp using the IRF510 have to deal with nominal RF issues, those specific to power FETs and specific limitations imposed by the IRF510. What makes it appealing is
> they are cheap and fairly easily had.
>
>
> Allison
>4576 2010-04-26 07:17:14 ajparent1 Re: problem with IRF510 push-pull PA 4579 2010-04-26 09:56:47 w4zcb Re: problem with IRF510 push-pull PA
http://www.radio-kits.co.uk/radio-related/20W_PA/20W_HF_PA_Construction_V2-1.pdf
built to compliment the PicaStar transceiver. Steve manages 60 dB of gain with 3 stages, I've built a half dozen of them over several years and never had a stability problem. The Ferrite is a little marginal4600 2010-04-28 10:28:32 g0kla Re: problem with IRF510 push-pull PA
In my case oscillation problems were in the radio rather than with the amplifier. I found that a 6dB pad on the input helped. (I was also probably driving it too hard.)
Once the amplifier was working I then had RF feedback issues in the radio because of the higher power. I had to add additional decoupling to the mic input, key line, mic amplfier etc.
It took me a while (several weeks) to realize that I had to fix the radio and not the amp. Once everything was RF tight, it seems to work well