|
Hello ? Now that both of my 3-500Zs have been de-gassed, I have a new question. ? ·???????? In the L7 (Drake) the bias is at 0 volts ·???????? In the Kenwood TL-522 it is -7.5 V as shown in the following diagram ? 
We read a bit of everything
? The objective is to put the L-7 into operation in SSB at medium power.
|
Hi Francois. ? In my experience, the 7.5V Zener provides a little bit of bias to limit the quiescent current of a pair of 3-500Zs running at around 3kV to 150mA. This is the requirement for Class AB2 linear operation. ? You could run them with Zero Bias (as was Eimac’s recommendation) to reduce circuit complexity, but the quiescent current would be much higher – for no improvement in linearity – hence the Zener diode. ? Most of us shy away from Zener diodes these days as a few decent size silicon diodes (1N5408 or 6A10) can be placed in series and provide the same bias voltage as the expensive Zener. ? Good luck with it. I do love all of my amps – 2 x 3-500Z on 6m, FU728F tetrode in Emtron DX-2SP for HF, GS35B triode in homebrew 2m amp for EME and YC156 triode for 160m/80m. All the triodes have silicon diodes providing the quiescent current bias voltage. ? Best 73, ? John VK6JX ?
toggle quoted message
Show quoted text
From: [email protected] <[email protected]> On Behalf Of Fran?ois via groups.io
Sent: Thursday, 27 February 2025 4:32 PM
To: [email protected]
Subject: Re: [ham-amplifiers] Grid bias 3-500Z? Hello ? Now that both of my 3-500Zs have been de-gassed, I have a new question. ? - In the L7 (Drake) the bias is at 0 volts
- In the Kenwood TL-522 it is -7.5 V as shown in the following diagram
? 
We read a bit of everything
? The objective is to put the L-7 into operation in SSB at medium power.
|
Yes. At 2.5kV plate volts, you probably don't need bias. At 3kV
in the TL-922, you definitely want bias but as VK6JX mentioned,
forward biased string of 6A10 diodes is a lot cheaper than a 50W
zener diode. You can place a 10uF to 20uF capacitor across the
string for better dynamic regultion as well.
73, Ale VK6APK.
On 27/02/2025 4:31 pm, Fran?ois via
groups.io wrote:
toggle quoted message
Show quoted text
Hello
?
Now
that both of my 3-500Zs have been de-gassed, I have a new
question.
?
·????????
In
the L7 (Drake) the bias is at 0 volts
·????????
In
the Kenwood TL-522 it is -7.5 V as shown in the following
diagram
?
We
read a bit of everything
?
The
objective is to put the L-7 into operation in SSB at medium
power.
|
What should happen if I add the red components in the attached schematic. I have the impression that it would do the opposite of the desired effect, we would block the tube on a strong RF input. ? I added the 100 k? for the cut-off of the tube in reception
|
John and Alek, my friend just spoke to Tibi at OM Power the other day. ?He called to order some FU728 tubes. ?Tibi said to him that his supplier of tubes in China is making a design change inside the tubes ( something to do with the grid structure) and the design change is now being implemented on the production line. He added that his next shipment will have the upgrades. ?We have found that at least 90% of the amps that come to our shops for repair are tube related. ?Just about all oh the OM4000’s present with at least one bad tube. ?Some long term owners of these amps have reported going through several tube changes already. ?Of course I don’t know how they use their amps and how many hours they put on them in a year but non the less it’s definitely been an issue. ?Hopefully if this is the case that there has been an engineering change, maybe this will be the fix. ?Remains to be seen. ?On another note, I have yet to use a Chinese GU74b so I can’t comment on them yet. ?Maybe some has?
On Thursday, February 27, 2025, 4:46 AM, John Sparkes via groups.io <vk6jx1@...> wrote: Hi Francois. ? In my experience, the 7.5V Zener provides a little bit of bias to limit the quiescent current of a pair of 3-500Zs running at around 3kV to 150mA. This is the requirement for Class AB2 linear operation. ? You could run them with Zero Bias (as was Eimac’s recommendation) to reduce circuit complexity, but the quiescent current would be much higher – for no improvement in linearity – hence the Zener diode. ? Most of us shy away from Zener diodes these days as a few decent size silicon diodes (1N5408 or 6A10) can be placed in series and provide the same bias voltage as the expensive Zener. ? Good luck with it. I do love all of my amps – 2 x 3-500Z on 6m, FU728F tetrode in Emtron DX-2SP for HF, GS35B triode in homebrew 2m amp for EME and YC156 triode for 160m/80m. All the triodes have silicon diodes providing the quiescent current bias voltage. ? Best 73, ? John VK6JX ? ? Hello ? Now that both of my 3-500Zs have been de-gassed, I have a new question. ? - In the L7 (Drake) the bias is at 0 volts
- In the Kenwood TL-522 it is -7.5 V as shown in the following diagram
? 
We read a bit of everything
? The objective is to put the L-7 into operation in SSB at medium power.
|
Francois,? ?with zero bias on the L4B, using Eimac 3-500Z's ( MU=130)? the? idle current is 220-230 ma = 575w to 600 watts of idle power.... = blast furnace.??
If using Chinese 3-500Z's ( MU=200).... idle current is a little less.?
I installed 10 x 1N5408's in series? ?for bias on my L4B's.? ? ?Idle current is now 100 ma with 2600 vdc....and just 40 ma with 1900 Vdc.??
Linearity with 1290 watts pep out on 75m is superb.? ?Linearity with 625 watts pep out (CW / 1900 vdc lower B+)? is also superb.?
I tested 1N5408's? with my lab supply and a resistor in series...on the bench...with 10 in series.? ?They run fairly warm with 1000 ma? continuous . On the Drake L4B / L7...on SSB it pulls 800 ma on peaks.....but only 400 ma average? DC plate current ( with processing).?
The bigger 6A10 diodes run warm? with 2000 ma? continuous.??
The bias issue is a dead simple fix.? Just wire 10 x diodes in series for the bias.? Forget the zener.? ?Zener's are not adjustable.? ?Zener's? won't handle surge currents like 1N5408 (200 amp surge) or 6A10 (400 amp surge) do.?
As you pull more current through the string of diodes, the V drop across each diode will increase slightly.? If you want to be a purist, install a? 500 uf to 1000 uf lytic across the entire string of diodes. Then the V drop is constant, and won't budge.?
IF using Chinese 3-500Z's ( MU=200),? 7-9 diodes will be required.? ?Install 10 x diodes.? ?Then it's easy to tap the diodes.? ?( shunt the last 1-2 diodes with a wire).?
Wire a .01uf @ 1 kv bypass cap? from each end of the diode string....to chassis.?
With the lower idle current, it will take a HUGE load OFF the power supply.? ?With the pair of 50K @ 50 watt wire wound bleeder resistors removed from the? B+ supply (and also the 5k @ 7 watt)? you will remove another? 63 watts from the power supply.?
Stock, keyed, but no drive applied, power consumed by the B+ supply was 663 watts.? ?After the bias diodes installed, and 50K bleeders removed,? ?power consumed by the B+ supply is just 260 watts. ( keyed, no drive applied).
663 watts - 260 watts =? ?403 watt? reduction in power from the HV supply.? ?That's a HUGE reduction in power.? ?The L4PS / L7PS top cabinet lid will then run cool to the touch.?
|
Hi Jim.
I tried a large value electrolytic across the bias string and
every time I keyed up, the ZSAC would take a while to stabilise at
the set level. When I stuck something like a 22uF in its place and
it seemed to be better. Without trying to work out time constants
and such, I figured that the 22uF would still provide the good
dynamic regulation during speech and CW transmissions. Dunno if I
did something wrong that initial time but for better or worse,
I've always settled on the lower value caps since then.
Cheers and thanks, Alek VK6APK.
On 28/02/2025 8:23 pm, Jim VE7RF via
groups.io wrote:
toggle quoted message
Show quoted text
Francois,? ?with zero bias on the L4B, using Eimac 3-500Z's (
MU=130)? the? idle current is 220-230 ma = 575w to 600 watts of
idle power.... = blast furnace.??
If using Chinese 3-500Z's ( MU=200).... idle current is a little
less.?
I installed 10 x 1N5408's in series? ?for bias on my L4B's.? ?
?Idle current is now 100 ma with 2600 vdc....and just 40 ma with
1900 Vdc.??
Linearity with 1290 watts pep out on 75m is superb.? ?Linearity
with 625 watts pep out (CW / 1900 vdc lower B+)? is also
superb.?
I tested 1N5408's? with my lab supply and a resistor in
series...on the bench...with 10 in series.? ?They run fairly
warm with 1000 ma? continuous . On the Drake L4B / L7...on SSB
it pulls 800 ma on peaks.....but only 400 ma average? DC plate
current ( with processing).?
The bigger 6A10 diodes run warm? with 2000 ma? continuous.??
The bias issue is a dead simple fix.? Just wire 10 x diodes in
series for the bias.? Forget the zener.? ?Zener's are not
adjustable.? ?Zener's? won't handle surge currents like 1N5408
(200 amp surge) or 6A10 (400 amp surge) do.?
As you pull more current through the string of diodes, the V
drop across each diode will increase slightly.? If you want to
be a purist, install a? 500 uf to 1000 uf lytic across the
entire string of diodes. Then the V drop is constant, and won't
budge.?
IF using Chinese 3-500Z's ( MU=200),? 7-9 diodes will be
required.? ?Install 10 x diodes.? ?Then it's easy to tap the
diodes.? ?( shunt the last 1-2 diodes with a wire).?
Wire a .01uf @ 1 kv bypass cap? from each end of the diode
string....to chassis.?
With the lower idle current, it will take a HUGE load OFF the
power supply.? ?With the pair of 50K @ 50 watt wire wound
bleeder resistors removed from the? B+ supply (and also the 5k @
7 watt)? you will remove another? 63 watts from the power
supply.?
Stock, keyed, but no drive applied, power consumed by the B+
supply was 663 watts.? ?After the bias diodes installed, and 50K
bleeders removed,? ?power consumed by the B+ supply is just 260
watts. ( keyed, no drive applied).
663 watts - 260 watts =? ?403 watt? reduction in power from the
HV supply.? ?That's a HUGE reduction in power.? ?The L4PS / L7PS
top cabinet lid will then run cool to the touch.?
|
On Sat, Mar 1, 2025 at 01:31 PM, Alek Petkovic wrote:
>Hi Jim.
>I tried a large value electrolytic across the bias string and every time I keyed up, the ZSAC would take a while to stabilize at the set level. When I stuck something like a 22uF in its place and it seemed to be better. >Without trying to work out time constants and such, I figured that the 22uF would still provide the good dynamic regulation during speech and CW transmissions. Dunno if I did something wrong that initial time but for >better or worse, I've always settled on the lower value caps since then.
>Cheers and thanks, Alek VK6APK.
##? what tube were u using ?? ?This makes no sense at all.? The string of diodes and the parallel? cap are in the circuit at all times...and in series with the 50-100K? cutoff resistor.? We used 7000 uf on buddy's 3x6 amp...and when 200 ma idle current drawn, it was rock solid.? ?I have also tried as much as 45,000 uf? (3" diam lytic with mounting bracket)....same deal...doesn't budge.? ?The V drop across the string of diodes is constant, whether the cutoff resistor is shunted or not.? This one baffles me.?
|
Hello ? Please do not think that I am not interested in your relevant information. I have to integrate them and, often, I redo the experiments not to verify their validity but to adapt them to my environment. ? In a previous theard, I showed how I was able to make the plate of a single tube red under a relatively low voltage with a very positive grid voltage. One of the next steps is to move to high operational voltages (CW and SSB) and again to be able to make the plate red in this new condition. ? The removal of the 50k + 50k +5k bleeder resistors worries me. 1.???? It quickly makes the high voltage disappear 2.??? It regulates the high voltage a little 3.??? It avoids stressing the amplifier with a no-load HT which will be equal to the peak voltage ? Here are the measured data: ? With the bleeder resistors and a zero anode current I note: ·???????? CW: 1500 V DC ·???????? SSB: 2500 V DC At the secondary of the transformer, depending on the transformation ratio and at no load: ·???????? CW: 230 x 2.95 = 678.5 V (960 V peak) ·???????? SSB: 230 x 4.14 = 952.2 V (1347 V peak) ? The high voltage at no load, without bleeder could be (estimated): ·???????? CW: 1920 V ·???????? SSB: 2694 V DC (we are therefore 200 V above the voltage without bleeder - Is this serious ? ? Concerning the polarization, I am a little in the dark. What you are proposing to me is a little like what I did a long time ago, in AM, with a "clamp" modulation on the screen of a tetrode. The idea was then to send the carrier only when the operator spoke. When he spoke weakly the carrier was not at full power. I did this by rectifying the BF to obtain a continuous voltage which was superimposed on the modulation on the screen. This artifice allowed the tube to be used at high power in the modulation peaks only. ? For these 3-500Z, I am trying to apply this principle. When the operator does not speak, the quiescent current allowing to switch from class B to class AB is not useful. ? ·???????? When the BF is weak, is it useful? the distortions in % will be higher but since the output power is low, is it serious. ·???????? When the BF is strong (PEP peak), it seems important to be clearly in class AB. Please confirm. ? Do you have a diagram showing where you place the diodes on the L-7? ? 73
|
Yeah mate. FIIK?.
Pair of 3-500Z and the cap was 4700uF/50V. When the amp was
keyed, the reed relay contact shorted the 15k resistor in series
with the diodes. The idle current would take a full second or so
to come up to the 150mA I wanted. The miniature reed relay failed
after many years and now there's a pink Opto22 job in its place.
I can test it again? soon, as I want to take my GS35B 6m amp out
of the rack for a modification. I'll stick a thumpin big cap in
there and see if I was dreaming all those years ago.
Cheers, Alek.
On 2/03/2025 12:10 am, Jim VE7RF via
groups.io wrote:
toggle quoted message
Show quoted text
On Sat, Mar 1, 2025 at 01:31 PM, Alek Petkovic wrote:
>Hi Jim.
>I tried a large value electrolytic across the bias
string and every time I keyed up, the ZSAC would take a
while to stabilize at the set level. When I stuck something
like a 22uF in its place and it seemed to be better.
>Without trying to work out time constants and such, I
figured that the 22uF would still provide the good dynamic
regulation during speech and CW transmissions. Dunno if I
did something wrong that initial time but for >better or
worse, I've always settled on the lower value caps since
then.
>Cheers and thanks, Alek VK6APK.
##? what tube were u using ?? ?This makes no sense at all.? The
string of diodes and the parallel? cap are in the circuit at all
times...and in series with the 50-100K? cutoff resistor.? We
used 7000 uf on buddy's 3x6 amp...and when 200 ma idle current
drawn, it was rock solid.? ?I have also tried as much as 45,000
uf? (3" diam lytic with mounting bracket)....same deal...doesn't
budge.? ?The V drop across the string of diodes is constant,
whether the cutoff resistor is shunted or not.?
This one baffles me.?
|
So I went to full voltage tests with the 2 tubes in place. With zero bias I measure ? ·???????? CW voltmeter 1700 V measures 236*8=1888 V intensity 175 mA; the tubes do not redden 298 W (2 tubes) ·???????? SSB voltmeter 2500 V measures 329*8=2632 V intensity 220 mA; the tubes redden 550 W (2 tubes) ? In low voltage tests 231 W (plate current 390 mA under 592 V) for one tube redden. The result therefore seems consistent. ? In my country (France) the power is limited to 500 W. ·???????? I could operate the L-7 on the CW position and by limiting the excitation 500 W RF output. ·???????? Periodically, run the amp in SSB position, without RF, with just the 220 mA quiescent current to make the tubes glow and thus maintain optimal degassing. ? I am concerned that by placing diodes in the cathode circuit to bias the grid slightly negatively, the quiescent current in SSB mode will not be sufficient
|
I think you are over thinking this degassing issue? By adjusting the amount of bias diodes you can adjust the anode current.
So I went to full voltage tests with the 2 tubes in place. With zero bias I measure ? ·???????? CW voltmeter 1700 V measures 236*8=1888 V intensity 175 mA; the tubes do not redden 298 W (2 tubes) ·???????? SSB voltmeter 2500 V measures 329*8=2632 V intensity 220 mA; the tubes redden 550 W (2 tubes) ? In low voltage tests 231 W (plate current 390 mA under 592 V) for one tube redden. The result therefore seems consistent. ? In my country (France) the power is limited to 500 W. ·???????? I could operate the L-7 on the CW position and by limiting the excitation 500 W RF output. ·???????? Periodically, run the amp in SSB position, without RF, with just the 220 mA quiescent current to make the tubes glow and thus maintain optimal degassing. ? I am concerned that by placing diodes in the cathode circuit to bias the grid slightly negatively, the quiescent current in SSB mode will not be sufficient
|
On Sat, Mar 1, 2025 at 05:48 PM, Fran?ois wrote:
?
The removal of the 50k + 50k +5k bleeder resistors worries me.
1.???? It quickly makes the high voltage disappear
2.??? It regulates the high voltage a little
3.??? It avoids stressing the amplifier with a no-load HT which will be equal to the peak voltage
?
Here are the measured data:
?
With the bleeder resistors and a zero anode current I note:
·???????? CW: 1500 V DC
·???????? SSB: 2500 V DC
At the secondary of the transformer, depending on the transformation ratio and at no load:
·???????? CW: 230 x 2.95 = 678.5 V (960 V peak)
·???????? SSB: 230 x 4.14 = 952.2 V (1347 V peak)
?
The high voltage at no load, without bleeder could be (estimated):
·???????? CW: 1920 V
·???????? SSB: 2694 V DC (we are therefore 200 V above the voltage without bleeder - Is this serious ?
Francois,? I have owned the? 1st L4B since 1977...and bought 3 x more since then.? ?The 1st? thing done to? any? L4B or L7? is to grnd the grids.? 2nd thing done is TOSS the 50k + 50K + 5k? 'bleeder'.? I was expecting the B+ to increase with their removal,? ?but the B+? DID NOT BUDGE !? ? 2650 vdc? on SSB.......and? 1900 vdc on CW...same as before.? ?Removal of the 50K+50K+5K had? NO effect.? ?Drake used the scheme simply to obtain +120 vdc / 90 vdc? for their ALC circuit.? There are NO ham amps that use bleeders.? They all use a? single 50k / 100k? across? each electrolytic.? ?Who cares how long it takes? ?for the B+? to drop to zero ?? ?There are a ton of screws on the top lid...+ a HV interlock. B+ comes down to zero after 1 min.? ?Plenty good enough.? Put 10 x 1N5408 / 6A10? /? PM-600? diodes in series? on some perf board.? ?It goes in series with the CT of the fil xfmr.? ?UN-banded end of the diode string faces the CT of the fil xfmr.? ?This will reduce the idle current down to 100 ma? on the 2650 position......and 40 ma on the 1900 position.? A? 50k to 100K? ?@ 2-3 watt? MOF resistor is wired in series with the new diode bias string.? ?Re-wire contacts of? the center pole of the TR relay? to shunt the resistor on? TX.? Done.? Jim? ?VE7RF
|
Put 10 x 1N5408 / 6A10? /? PM-600? diodes in series? on some perf board.? ?It goes in series with the CT of the fil xfmr.? ?UN-banded end of the diode string faces the CT of the fil xfmr.? ?This will reduce the idle current down to 100 ma? on the 2650 position......and 40 ma on the 1900 position.?
A? 50k to 100K? ?@ 2-3 watt? MOF resistor is wired in series with the new diode bias string.? ?Re-wire contacts of? the center pole of the TR relay? to shunt the resistor on? TX.?
Could you make me a diagram please? -- F1AMM Fran?ois
|
Can one of u folks pse help me out here, as I have no clue how to post a simple diagram on here.?
|
Draw it and take a picture and post that????
Can one of u folks pse help me out here, as I have no clue how to post a simple diagram on here.?
|
Hello ? Okay, I got it. It's the receive polarization that I didn't understand. Now it's ok. ? What quiescent currents should I get to and in what mode (SSB/CW), knowing that I want to output 500 W PEP?
|
Hi Francois.
I would suggest you leave it as zero bias, without any diodes. As
you say, with just 500W PEP output, the tubes will not get hot
enough to de-gas.
Your solution of running the amp at 220mA without drive is very
good and it will ensure that your tubes are always healthy.
So keep it as it is and don't change the bias. It will work fine
for your application.
73, Alek VK6APK.
On 3/03/2025 11:39 pm, Fran?ois via
groups.io wrote:
toggle quoted message
Show quoted text
So
I went to full voltage tests with the 2 tubes in place. With
zero bias I measure
?
·????????
CW
voltmeter 1700 V measures 236*8=1888 V intensity 175 mA; the
tubes do not redden 298 W (2 tubes)
·????????
SSB
voltmeter 2500 V measures 329*8=2632 V intensity 220 mA; the
tubes redden 550 W (2 tubes)
?
In
low voltage tests 231 W (plate current 390 mA under 592 V)
for one tube redden. The result therefore seems consistent.
?
In
my country (France) the power is limited to 500 W.
·????????
I
could operate the L-7 on the CW position and by limiting the
excitation 500 W RF output.
·????????
Periodically,
run the amp in SSB position, without RF, with just
the 220 mA quiescent current to make the tubes glow and thus
maintain optimal degassing.
?
I
am concerned that by placing diodes in the cathode circuit
to bias the grid slightly negatively, the quiescent current
in SSB mode will not be sufficient
|
Hello ? Thanks Alek for resuming the dialogue. You are the only one who has reported this risk of non-degassing to me. Like me, you are faced with a reduced regulatory RF power ? You had told me /g/ham-amplifiers/message/40898 that the gettering was done by the anodes. If you run 500W PEP only, your tubes will quickly become gassy. They have no separate getters. The gettering is done by the anodes and so they must become red during operation to keep them healthy. This happens in Australia as well where we have 400W PEP limit. At that level, the anodes never get red, so the tubes fail very quickly. You will need to exercise your amplifier at full power regularly to ensure healthy tubes. This risk does not seem to worry the other participants at all. ? Your point is confirmed in this document https://www.w8ji.com/Amplifiers.htm The fastest way to ruin a 3-500Z or other glass electric grid tube is to never expose the anode color for an extended period of time! Storing a 3-500Z tube for many years without using it almost guarantees that a breakdown will occur the first time a high peak voltage is applied to the anode. Since the tubes had not been used for 15 or 20 years, I did not want to expose them to the risk of arcing due to imperfect degassing. So I came up with a solution, which I have outlined, to make the plates glow under a relatively low anode voltage by applying a positive voltage to the grid (actually positive to the cathode). ? I proposed ·???????? To use the SSB position 2500 V 220 mA only to periodically degas the tubes and this without RF ·???????? To use the CW position 1700 V 175 mA during real SSB traffic with RF=500W ? Eventually, I can find a solution so that in CW position the anode current is lower. De la part de Alek Petkovic via groups.io
贰苍惫辞测é?: mardi 4 mars 2025 10:43
|
