|
Hello all,
?
I'm finally finding time to assemble my QMX+. I already have a stockpile of TN0110 FETs from my QDX reworks. Wondering if I should try TN0110 on the QMX+ or stick to the BS170s. My thinking:
- QMX/QMX+ are not as prone to BS170 failure as QDX. This may be due to no L14 kickback issue because of envelope shaping. Or it might be another reason. But empirically I see fewer reported failures.
- On QDX it seemed that switching to TN0110 resulted in slightly higher power output, but on QMX people who tried TN0110 reported similar or slightly lower power out. (Except on the high-bands?)
- Because QMX/QMX+ has the switching mode 5V and 3.3V supplies, it seems more reasonable to run a unit built for 12V higher than 12V to get a little bit more from the TN0110s. In the QDX running higher than 12V seemed likely to stress the stock regulators, which already ran pretty hot. Not that I would necessarily do this, but I'm curious what people think.
- In the QMX, there's basically no room to make further mods to accommodate the reversed pins of the TN0110, while in the QMX+ there is plenty of room to add a heatsink to the top flat side and/or add fans. (In the QDX there is just enough room to put in a and I've been operating happily with this setup for a while now.)
On balance I would plan to just put the BS170s in for now, but I'm interested what others think in general, or specifically on any of the (possibly erroneous) thoughts above. Thanks!
|
Hi Peter,
All other things being equal, TN0110 is a better switching device than BS170. First, its 100V maximum rating exceeds the BS170’s 60V maximum rating so if you are brave enough to try operating the QMX+ above 12V (not necessarily recommending this), it won’t be destroyed by overvoltage. Current *may* be a different matter but see below.
Second, TN0110 has a lower turn-on threshold than BS170 which means it’s a better match to the relatively low 5V switching waveform applied to the gate. To achieve the same level of conduction in a BS170, you’d need at least a 6V switching waveform which, unfortunately, isn’t available. This lower threshold allows the TN0110 to generate slightly more power on all bands than BS170. TN0110 has a lower channel resistance than BS170 at the same gate drive level so it sources more current. Depending on the overall channel resistance vs drain current, it may run slightly hotter or cooler. It’s hard to know which without a direct comparison of both devices.
Since you have more room for a better heat-sinking methodology in QMX+, heat-sink away with the proper application of thermal grease and a nice copper-finned or aluminum-finned heat sink (copper is better).?
Overall, I think you’re in great shape to use TN0110 and I don’t see any downside. If you decide to go this route, please do report back your experiences here. And please snap some pictures of your heat-sinking strategy.
Tony AC9QY
Hello all,
?
I'm finally finding time to assemble my QMX+. I already have a stockpile of TN0110 FETs from my QDX reworks. Wondering if I should try TN0110 on the QMX+ or stick to the BS170s. My thinking:
- QMX/QMX+ are not as prone to BS170 failure as QDX. This may be due to no L14 kickback issue because of envelope shaping. Or it might be another reason. But empirically I see fewer reported failures.
- On QDX it seemed that switching to TN0110 resulted in slightly higher power output, but on QMX people who tried TN0110 reported similar or slightly lower power out. (Except on the high-bands?)
- Because QMX/QMX+ has the switching mode 5V and 3.3V supplies, it seems more reasonable to run a unit built for 12V higher than 12V to get a little bit more from the TN0110s. In the QDX running higher than 12V seemed likely to stress the stock regulators, which already ran pretty hot. Not that I would necessarily do this, but I'm curious what people think.
- In the QMX, there's basically no room to make further mods to accommodate the reversed pins of the TN0110, while in the QMX+ there is plenty of room to add a heatsink to the top flat side and/or add fans. (In the QDX there is just enough room to put in a and I've been operating happily with this setup for a while now.)
On balance I would plan to just put the BS170s in for now, but I'm interested what others think in general, or specifically on any of the (possibly erroneous) thoughts above. Thanks!
|
Interesting....
I hadn't considered changing output devices on any of my Q*X's as I expected Hans already put much effort into selecting the best part for the job.?
?
A quick look at the datasheet for the TN0110 shows it has much higher input capacitance than the BS170. Does that not impact its performance in this role, given the available drive?
?
?
|
Having in mind future SSB implementation in QMX/+ there is a risk the SSB generation software will be optimized for BS170 characteristics. Doing the mod for fun for myself - ok. It can be always reversed.?
73!
Pawe?
|
Hi Ronan ?
I hadn't considered changing output devices on any of my Q*X's as I expected Hans already put much effort into selecting the best part for the job.?
Thanks Ronan but you are too kind ;-)? ?I am constrained crucially by cost and availability considerations. An individual constructor may not be so constrained; and for this reason it is not necessarily?the case that no "upgrade" of a particular part is possible! For example?- if you buy a Ford Pickup truck and you really love it, and you want to pimp it up, and you decide to give it a custom gold plated finish, leather upholstery etc - then it will look very nice and fine for you that you did it. But Ford sure as hell aren't going to do offer it as standard or probably not even as an option, because it would not suit their cost and availability constraints.??
? A quick look at the datasheet for the TN0110 shows it has much higher input capacitance than the BS170. Does that not impact its performance in this role, given the available drive?
I do think it can do, and that would impact the power output and efficiency at higher frequencies. I am not sure if anyone here has tested that, apologies if so and I have missed or forgotten it.?
Aside from the higher gate capacitance, I think there are other differences too which may or may not impact the suitability. Datasheets are often difficult to interpret and particularly to compare between parts from different manufacturers, because often the specified data is at a certain operating point which is different between the two datasheets. But I would make the following observations: - BS170 is rated 500mA continuous, TN0110 only 350mA.?
- TN0110 has a higher rated power dissipation 1W vs the BS170's 830mW, and also a lower thermal resistance 132 C/W compared to the BS170 150 C/W. These are advantages for the TN0110 but this is also typically what you see: a device capable of handling more watts typically has higher capacitances.?
- BS170 has a lower typical ON state resistance than TN0110 which would provide some slight efficiency advantage
I should also note that I'm not really yet convinced by the proposition that the lower gate threshold voltage of the TN0110 is an advantage. The gate threshold is specified at Vds = Vgs and Id = 1mA; BS170 min-typ-max is 0.6-2.1-3V and TN0110 is min-max 0.6-2V (at the same operating point). However these operating conditions are not really similar to what is really encountered in the circuit and I think it is more correct to look at the Saturation Characteristics performance curve (which in the BS170 sheet is labelled "On-Region characteristics"). Taking into account the difference in scales of the graphs (due to the fact TN0110 is in some way a "bigger" device), it's hard for me to see any difference between the two transistors, at the 0-5V squarewave drive and expected Drain-Source currents in the actual circuit. In both cases a 5V square wave should saturate the transistor in the ON/OFF states. I don't see an easy way from the datasheets to say which transistor would be best. I may be wrong.
Personally and again I could be wrong, I actually like a slightly higher threshold, IF it still allows full saturation at the "high" of the drive circuit waveform in the application, because it appears to me to be closer to symmetry. The rise/fall time of the logic gate IC that drives the MOSFET gate is not infinitely fast. It appears to me that if the gate threshold was optimally at the mid-point (2.5V here), that would give the best symmetry in the transistor switching times; ideally you want as close to 50% timing as possible, as any deviation from that would I think start to cause overlap cancellation in the output transformer which will impair efficiency. To imagine in extremes, that the transistors are ON for 75% of the time and OFF for 25% of the time; then there is a considerable overlap in the push-pull circuit where BOTH the push side and the pull side are ON, and fighting each other via the output transformer. So as close as possible to 50% should be desirable and maximize efficiency.?
As always the proof of the pudding is in the eating, not in the studying of datasheets. But the pudding should always be cooked in the same kitchen, by the same chef, and eaten by the same customer making a decision, under the same conditions (not more hungry one day, than the next). In other words it has to be ONE QMX, all the LPFs exactly the same, the same person building it, the same measurement equipment under as identical conditions as possible. One person on one side of the planet can't build is QMX one way, measure it with his equipment in his lab, and compare results meaningfully with another person on the other side of the planet who used TN0110 and measured in his own lab. There are in that case just too many variables changing.? ?
|
How did you know about my car? That's just creepy ?
|
Hi,
I was using 4 x TN0110 in my Lo bands QMX. What I got was a good safety feeling when operating it at +14V instead of 12V and consequently a bit higher output RF power. I would and am not using them in the QMX+.
--
73 Bojan S53DZ
|
How did you know about my car? That's just creepy ?
?? I know where you live, too. So don't even think about trying anything???
|
I have built up my 9V-QMX 60-15m with TN0110 - the rounded side was ground flat, the top side was provided with an additional cooling plate, LPF-coils optimized.?
The information comes from the terminal control display (current from the power supply, temperature sensor)
Operating voltage 9.6V, at the finals 8.2V
Current consumption 0.9A (17m) to 1.1A (20m)
Output between 5.2W (60m) and 5W (20m), 18m 4W, 15m 4.3W
Power loss at the finals 1.7W (60m) to 3.33W (15m) increasing. (= ((IinTX-IinRX)x8,2V)-Pout)
So far I have not seen any clear differences to the BS170
My temperature test for 50°C of the cooling surface (the permissible Id and the power dissipation drop to 80%) at 25°C room temperature:
In CW no 50°C are reached with a CQ loop.
In FT8 no 50°C are reached with CQ loop from 60-20m, but on 18MHz after about 14x CQ and on 21MHz after 7xCQ
A test with Olivia (Fldigi, passes up to 2min) on 17m was aborted immediately due to the rapid temperature increase caused by the long transmission cycles.
Test with TinySA: Lowest harmonic attenuation -48dB on 60m, otherwise at least -56dB, sensitivity between -118dB (21MHz) and -128dB (7MHz) at approx. 10dB S/N, CW
I am satisfied with my QMX, I use it mainly for CW and some FT8
73 Reiner
|
The Xeigu G90 uses RD16HHF1 (Mitsubishi) and I need to order some (ours failed) but in this context have these devices been considered as PA for any of Han's excellent designs??
?I know it is a very complex study to determine cross compatibility (beyond my pay-grade) but these devices seem to have favorable specifications for our applications here (Q*X-cool awesome rigs).
I totally understand and appreciate Hans' prior responses-- makes total sense (I've attempted kitting- PIA) but maybe some of you out there who like to push the boundaries and leather clad your gold upholstered Fords (;-)) might share your thoughts &/or experiences.
'73 AD6XJ ?
|
The driver will not open the RD. Need a level converter
--
73, Oleh, UY2ZA
|
Hi Robert,
?
The RD16HHF1 is EOL, so using it would not be a wise move on Hans' (or anyone's) part!
?
73,
?
Robert, WA2T
|
And what do you think of this WB2CBA modification from BS170 mosfets to FDT86256 PA mosfets? I bought the kit for my QDX but never installed it as I take always high precautions with the tuning of my antenna with a NanoVNA before transmitting.?
73 - Pierre - FK8IH
|
Both are 3X more expensive than BS170 at DigiKey. I have used my stock QMX on FT8 and. CW for more than a year with no failures. I expect my QMX+ to do the same. I always have SWR protection on and check SWR before operation. What is the benefit to make the switch?
On Sat, Jul 20, 2024 at 6:08?PM Pierre FK8IH via <jb.gallauziaux= [email protected]> wrote: And what do you think of this WB2CBA modification from BS170 mosfets to FDT86256 PA mosfets? I bought the kit for my QDX but never installed it as I take always high precautions with the tuning of my antenna with a NanoVNA before transmitting.?
73 - Pierre - FK8IH
|
Actually? NOT EOL? ,? another mfr has taken over the RD16HH and similar line.? My contact has shown me photos of the new devices
I don't think it's the WhoFLUNGDUNG company operating out of a tin Shack on the upper? Bejing river either .THese perform precisely as per the original specificantion.? They just don't hve? the mitsub logo on them?and the lettering is straighter.? I would not be surprised if the mafr is some sort of subsidiary of Nxp who have the IP on the MRF101A and B.?
The holding company of this new mafr is allegedly Japanese.? It's weird how the origin of chips and RF silicon is via? smoke and mirrors and these high tech devices? get mafr'dout in never never land and get filterd by the likes of? Mouser ; so you have no idea of the provenance.?
The chap with the inside running? (who seems to import RF devices in bulk)? hangs out on the? Friends of? Toroid King site or whatever name it is... If I look it up I'm sure to lose this edit.? He has lots of letters post his name? (possibly some of his? own invention)? and rents out some house on the US Virgin Islands.? That narrows the field.?
?
?
TEF
|
dl8lrz Reiner
Pierre - FK8IH
Timothy Fidler