开云体育

Hi,

?

Would it be difficult to change BS170 final transistors to LDMOS ones??

?

Regards

Hi Mateusz,

Gate input capacitance and output capacitance of any replacement transistor must be in the same neighborhood as the BS170. The closest match I’ve found is TN0110 which some here have been using with success. It’s a 100V device as opposed to BS170 which is a 60V device and provides better protection against high SWR voltages. The only thing to know is TN0110 must be installed with the flat side up to account for the reversed pinout.

Tony AC9QY

Did not investigate the component TN0110 further, but may provide opportunity in combination with an "U" shaped heatsink instead of a washer when installed in a neat way.

?

Harold PE1RJP - Happy QMX+ owner

No it would just be difficult to make it work?

As Harold mentions, it can be done with TN0110s, I've done so in several QDXs. But I would NOT do this in a QMX.

?

?

-Nate

N8BTR

It makes more sense to make the change on a QMX+ since on 80m and 160m the peak voltage, even with 1:1 SWR exceeds the BS170 maximum voltage rating.

Hi Chris,

?

I am not sure why there is a difference of peak voltage on the lower frequency bands unless it is due to higher gain at lower frequencies.?

I would be hesitant the to change the finals in the QMX line of radios, as Hans has developed a CESSB algorithm that is based on the BS170 characteristics. ?The QDX and QCX do not have this restriction.?

Evan

I did not really figure out why the peak voltage gets worse at lower frequencies.? In general these peak voltages are caused by the harmonics reflected back into PA transistors by the LPF.? It could be how the higher harmonics stack up because of the particular phase of the reflected higher harmonics.??

?

The plot shows peak voltage vs SWR for 40m to 160m bands

Chris,

?

I am not sure I am reading the chart correctly, but it looks like there is a significant power output difference for the different bands. ?The line color for band is not in the picture.

?

Did you record output power for each band?

?

73

Evan

AC9TU

The plot does not address power. Only peak voltage across the BS170s. However the 160m has higher power than the other bands.

You can see that 160m and 80m exceed the 60V ratings of BS170 and are marginal on 60m and 40m.

?

When evaluating the effects of ATU switching (used my Elecraft T1) I was able to capture ~100v very fast spikes.? Breakdown from these spikes may cause cumulative damage and ultimately total failure.

?

?

FYI?

On my plot I should of suppressed "0" and "-1" on the x axis?

Chris,

?

I guess I need new glasses, as I now see the band designation in the prior posts.? Looking at the SWR of 1, the measured voltage does seem to follow projected power as frequency changes.? A proper test is to keep the power out at 5 watts for each band at 1:1 SWR, adjusting the voltage to the radio to maintain the correct power level.

?

The PA's class D design power output is significantly impacted by the impedance of the output stage, which includes the filter and the antenna.

?

The spike at the end of transmission for the QDX is well-documented.? The QDX has no key-shaping circuit to prevent the abrupt end.? I have measured it at well over 120 volts, which is peak voltage, not peak-to-peak.? A commutating diode fix across L14 removes this spike (similar to diode protection for a relay).

?

The QMX does have a key-shaping circuit, and I know of no spikes on key up that have been detected.? If you are seeing spikes, a solder bridge in the Q507 circuit could keep it on all the time.? This happened to one of my QMX builds.? Q507 is the key shaping circuit.

?

73

Evan

AC9TU

Chris,

It’s clear that the BS170 maximum Vds rating is greatly exceeded on 80 and 160 meters and this is indeed due to higher gain in the FETs at lower frequencies. I saw this gain difference vs frequency in the sims I ran using JZ’s model long ago. I proposed TN0110 as an alternative to the BS170 back then to handle the voltage spikes in the QDX. QMX fixed the spikes but that was prior to opening up the 80 and 160 bands where high voltage is now showing up for an entirely different reason. I know Hans planned to limit the drain voltage in these two bands but I don’t know if he completed this effort. From your graphs, it appears that this feature isn’t working in whatever firmware version you have. In any event, lowering the supply voltage when operating on 80 or 160 should lower the PA excursions but they might still be exceeded to some degree. Have you tried running your tests at a lower supply voltage?

Tony

Good, had me puzzled.

So where the bottom axis now reads:

-7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

it should read:

-9 -8 -7 -6 -5 -4 -3 -2 1 2 3 4 5 6 7 8 9 10 11

?

An informative graph, thanks!

?

Do those SWR's all assume a purely resistive load?

The load could also be capacitive or inductive, which would also affect the peak voltage.

?

Keeping the SWR below 2 is clearly a good idea on 80m and 160m.

If you have a VNA so you can see more than just the SWR, shoot for an impedance under 50 ohms.

?

Tony's idea of substituting TN0110's could be good for those not planning to use the new SSB firmware.

Perhaps Hans can someday have a menu option to allow SSB mode using the TN0110's.

?

Going to TN0110's adds another $4 to the parts cost, I'd guess the kits will continue going out with BS170's.

Using Q507 as a linear regulator to reduce the supply voltage to the? BS170 finals on 80m and 160m should work

or you could vary the power supply external to the rig.

?

A future kit might be able to use an SMPS instead of Q507 to keep efficiency high.

But if Hans succeeds in getting SSB firmware out with the hardware as it is, that's a huge win.

And time for him to celebrate with a 6 month break in some exotic foreign land.

?

Jerry, KE7ER

?

?

On Fri, Jan 3, 2025 at 12:27 PM, Chris KB1NLW wrote:

On Fri, Jan 3, 2025, at 04:41 PM, Jerry Gaffke wrote:

A future kit might be able to use an SMPS instead of Q507 to keep efficiency high.

But if Hans succeeds in getting SSB firmware out with the hardware as it is, that's a huge win.

And time for him to celebrate with a 6 month break in some exotic foreign land.

?

On Fri, Jan 3, 2025 at 02:41 PM, Jerry Gaffke wrote:

Tony's idea of substituting TN0110's could be good for those not planning to use the new SSB firmware.

Perhaps Hans can someday have a menu option to allow SSB mode using the TN0110's.

?

Why would different FETs affect the new SSB firmware?? I understand that the phase and amplitude of the drive the the PA FET have different routes.

?

One possibility is a change in the propagation delay from the MS5351 to the PA output but that's going to be of the order of nanoseconds.? Would that be enough to affect things?

The other would be a change in the relationship between the PA voltage and the RF amplitude wouldn't it?? Would that be more of an issue?? There's linearity and gain to think of.? Or is there something else?

?

I stress that I don't know, but several of you are saying confidently that changes in this area WILL make a difference.? You obviously know more about this than me and I'm curious about what it is that a different PA FET type would change that would make the new SSB not work as well.

?

Chris, G5CTH

On Sat, Jan 4, 2025 at 01:51 PM, Chris wrote:

Why would different FETs affect the new SSB firmware??

I will try and dig the post out but Hans explained this in some detail on one of the threads when discussing the SSB implementation.? From memory something to with the CESSB being tuned to the characteristics of the BS170 and being quite specific to them.

On Sat, Jan 4, 2025 at 05:51 AM, Chris wrote:

?

Basically, capacitances inside the BS170's change significantly with changes in the supply voltage from Q507.? Those capacitances cause delays of the RF going through the BS170's, and this significantly affects the RF phase.? Hans models those delays in firmware to avoid distortion in the SSB signal.? He also said in a later post that variations in particular BS170's were not enough to significantly distort the signal.? I don't think he has experimented with TN0110's, if we're really lucky perhaps the SSB signal will still be usable.

?

?

?

Almost all HiFi audio systems built today use class D power amps.? They can be very low noise.? The class D audio amp has exactly the same topology as a buck SMPS, but curiously the power can flow in the reverse direction as a boost mode SMPS, taking power from the load and pushing it back into the main power supply's filter caps.? A new concept for me.? However, since Q507 cannot pull power from the BS170's, I doubt anything more than a standard SMPS is needed here.? Since the modulator need only respond fast enough to keep up with changes in amplitude at a few KHz, a buck SMPS operating at a couple MHz should have no trouble with audio fidelity.

?

Jerry, KE7ER