|
Hi all
A small number (approx 0.1% of this group) are experimenting with alternative PA transistors and methods to improve the robustness of QRP Labs PAs. I applaud their efforts, reading their?adventures with?great interest, and we should always encourage and admire experimentation and improvement.?
However I want to re-assure the remaining 99.9% of you that the BS170 is an excellent choice, reliable, durable, robust and inexpensive. I believe it is difficult or impossible to find a better transistor today if you include availability, price, packaging and characteristics as criteria.
BS170 MOSFETs have been used in QRP transceivers for 20+ years. For very good reasons. QRP Labs has been using them for 10+ years, first in the Ultimate-series QRSS/WSPR transmitters then in the transceiver kits starting with QCX. To date over 20,000 QCX transceivers sold have three BS170s in parallel Class-E, and almost 10,000 QDX/QMX transceivers sold have the quad of push-pull BS170's in the Class-D design. Years of experience have shown the BS170 largely "just works", reliably, day in day out for years.?
Generally when it comes to MOSFETs, the bigger the little lump of silicon in the middle, the more current it can handle, and?watts it can dissipate, but the higher the capacitances involved, which requires lower impedance drive particularly at higher frequencies. There's no free lunch. "Real" RF transistors may have little tweaks to improve this or that parameter or characteristic curve shape. In a Class-E or Class-D PA, hard switching is all we need anyway, and again the BS170 is great for this.?
The BS170 lies at a particular sweet point. The specifications of the device are just right for our QRP efforts at HF. The QDX can even produce over 2 Watts right up on 4m band (70 MHz). The 500mA current rating and 833mW dissipation rating make them well suited to the 5W PAs of QCX, QDX and QMX transceivers.?
Very importantly, they are low cost, easily available everywhere, and the TO92 through-hole packaging makes them perfect for that large majority of us that don't have equipment and/or inclination and/or experience for SMD soldering, in those situations where transistor replacement is needed.?
Yes, indeed there are circumstances where they could fail sometimes. There are people who push their?transceivers for absolute maximum power output, run with higher battery voltages, or perhaps SWR mismatches. I personally don't recall having lost a QCX/QDX/QMX PA except for that night there was a big storm and I was too lazy to get out of bed and unplug the antenna, and static zapped them. But in that case, it's easy and cheap to replace them.?
I don't see anything much not to love about a few BS170s working hard, when you want 5W at HF. If you want more power, don't waste time going for 6 or 7W. Do it properly and go for at least a 5x or 10x increase in power. For that matter the IRF510 comes next, a whole new praise topic, again easy, robust, reliable, inexpensive, and commonly available ;-)?
|
Hans, thank you for your encouragement, and I actually agree with what you said about BS170. The same can be said about a lot of other devices that hams have used historically "off label" but quite effectively. Many versions of 6L6 transmitters were popular, but 6L6 wasn't sold for that. A lot of television receiver set tubes were popular in amateur radio rigs. Then, when CB became popular in the US, Mitsubishi, Sanyo, Toshiba, and others made many RF transistors for 4W at 30MHz, but some were usable for 10W even at 54MHz, and we adopted whatever could be used effectively. BS170 is along a similar line, and it works.
But some of us keep that tradition. We have LDMOS for handheld mobile radios, cellular phones, cable television, satellite radio, wireless network equipment, ISM, etc., available and relatively inexpensive. Those tiny LDMOS transistors may still cost 10x or 50x BS170, so the cost would not fit well with sub-USD100 kits, but it is within reach of some user mods.
The biggest obstacle is that those new devices were almost never sold in through-hole packages. Some LDMOS transistors were sold in TO-220, but those are largely obsolete and expensive now (and many counterfeits sold). Besides, they are overkill for 5W. Tiny 3W transistors come in SOT-89; bigger ones are DFN or PLD packages. As I said in other threads, the TO-92 package is appropriate for kits for the enjoyable building experience for the majority of builders. But once again, that is within reach of some user mods.
What if an LDMOS were available in TO-92 for a few cents? Would that be adapted among amateurs? I absolutely think so. But that does not mean I am criticizing BS170. I've used and abused a lot of different transistors, including BS170. Neither did I criticize 6L6, 6CL6, 12BY7A, CB final transistors, etc. I'm simply looking at all options for my user mod where the cost and SMD package are less of an issue than the standard kit design.
|
Well stated Hans !
Jim WB4ILP?
|
Hans,
As one of the 0.1% thank you for the encouragement to experiment on QDX. I was fearing to upset you with my QDX PA mod though it¡¯s just an innocent alternative PA test on QDX just to see what will be the outcome. Then I shared my experimentation with community as I always do.
As a fan of BS170 and using it in all my own rigs I can¡¯t agree more of your thoughts on BS170. You are absolutely right.
73
Barb WB2CBA
?
|
I¡¯m currently working on a UL3s for 10 meters and decided to solder in BS170s in the remaining slots and the output power actually went way down. ?Is this due to parallel capacitance adding up to the point that it is no longer tuned well for 10 meters? If I want to get to 3 or 4 watts with it is my best option to add an external amp? If necessary Ive got several old CB radios that I can hack and drive the driver amp with a single BS170. I¡¯ve also got a few IRF510 and IRF511 around, the latter which is out of production.
Thanks,
toggle quoted message
Show quoted text
On Sep 22, 2023, at 8:02 AM, wb2cba@... wrote:
? Hans,
As one of the 0.1% thank you for the encouragement to experiment on QDX. I was fearing to upset you with my QDX PA mod though it¡¯s just an innocent alternative PA test on QDX just to see what will be the outcome. Then I shared my experimentation with community as I always do.
As a fan of BS170 and using it in all my own rigs I can¡¯t agree more of your thoughts on BS170. You are absolutely right.
73
Barb WB2CBA
?
|
Jerry, maybe you meant to post this on a different thread?
In general and short, CB final transistors work well for 3-5W HF transmitters and give about 10dB gain. One good advantage of some (but not all) CB final transistors were designed for very deep AM modulation so they can take quite high collector voltage for the power level. Those are suitable for class C/D/E. The downside is that they were discontinued decades ago and ebay, Amazon, etc. are full of counterfeits (many not even RF transistors) so they are not good for new design. But if you already have some, it should work.
|
On Fri, Sep 22, 2023 at 12:44 PM, Hans Summers wrote:
To date over 20,000 QCX transceivers sold have three BS170s in parallel Class-E, and almost 10,000 QDX/QMX transceivers sold have the quad of push-pull BS170's in the Class-D design. Years of experience have shown the BS170 largely "just works", reliably, day in day out for years.?
Amen. 73s Ted 2E0THH
|
Ryujii,
My apologies if my reply was off topic. Maybe it was only on-topic for me :-). ?I thought it was about the merits of the BS170 and my intention was to point out an apparent frequency limit that I¡¯ve run into at 10 meters insofar as increasing power output by adding more BS170s. ?So, as a result, I¡¯m looking for alternative ideas to increasing power just a bit.
73, Jerry, AC5JM?
toggle quoted message
Show quoted text
On Sep 22, 2023, at 9:30 AM, Ted 2E0THH <qrp@...> wrote:
?On Fri, Sep 22, 2023 at 12:44 PM, Hans Summers wrote:
To date over 20,000 QCX transceivers sold have three BS170s in parallel Class-E, and almost 10,000 QDX/QMX transceivers sold have the quad of push-pull BS170's in the Class-D design. Years of experience have shown the BS170 largely "just works", reliably, day in day out for years.?
Amen. 73s Ted 2E0THH
|
Ok Jerry - I thought this thread was Hans expressing his viewpoint and various tradeoffs made as the kit designer. Obviously, those who are doing user-level modifications have greater freedom in terms of cost and other technical constraints as long as we can take the risks. Regarding freq response, class C/D/E/F have more stringent requirements in dynamic characteristics than class AB and it can be tough on high bands. But I don't think anyone is promoting BS170 as a universally good choice for final transistors, either.
|
Jerry wrote: I¡¯m currently working on a UL3s for 10 meters and decided to solder in BS170s in the remaining slots and the output power actually went way down. Is this due to parallel capacitance adding up to the point that it is no longer tuned well for 10 meters? ... Jerry, here is a snap of the datasheet for the Si5351 used in the UL3s. Note the spec for maximum output capacitance (15pF). One BS170 (according to its datasheet) is already 24pF-40pF. Couple that with a turn-on threshold voltage that is also high (3V maximum) relative to a 4.5 V drive up-level and I think then the answer to your question is "yes". JZ On Fri, Sep 22, 2023 at 10:56?AM Jerry Moyer via groups.io <ac5jm@...> wrote:
I¡¯m currently working on a UL3s for 10 meters and decided to solder in BS170s in the remaining slots and the output power actually went way down. Is this due to parallel capacitance adding up to the point that it is no longer tuned well for 10 meters? If I want to get to 3 or 4 watts with it is my best option to add an external amp? If necessary Ive got several old CB radios that I can hack and drive the driver amp with a single BS170. I¡¯ve also got a few IRF510 and IRF511 around, the latter which is out of production.
Thanks,
Jerry AC5JM
On Sep 22, 2023, at 8:02 AM, wb2cba@... wrote:
?
Hans,
As one of the 0.1% thank you for the encouragement to experiment on QDX. I was fearing to upset you with my QDX PA mod though it¡¯s just an innocent alternative PA test on QDX just to see what will be the outcome. Then I shared my experimentation with community as I always do.
As a fan of BS170 and using it in all my own rigs I can¡¯t agree more of your thoughts on BS170. You are absolutely right.
73
Barb WB2CBA
|
One correction. That capacitance is not a maximum spec, but rather a
test condition. My mistake. Still, I think the total capacitance may
be too much for the chip to swing at 28 MHz.
JZ
toggle quoted message
Show quoted text
On Fri, Sep 22, 2023 at 1:32?PM John Zbrozek <jdzbrozek@...> wrote:
Jerry wrote:
I¡¯m currently working on a UL3s for 10 meters and decided to solder in
BS170s in the remaining slots and the output power actually went way
down. Is this due to parallel capacitance adding up to the point that
it is no longer tuned well for 10 meters?
...
Jerry, here is a snap of the datasheet for the Si5351 used in the
UL3s. Note the spec for maximum output capacitance (15pF). One BS170
(according to its datasheet) is already 24pF-40pF. Couple that with a
turn-on threshold voltage that is also high (3V maximum) relative to a
4.5 V drive up-level and I think then the answer to your question is
"yes".
JZ
On Fri, Sep 22, 2023 at 10:56?AM Jerry Moyer via groups.io
<ac5jm@...> wrote:
I¡¯m currently working on a UL3s for 10 meters and decided to solder in BS170s in the remaining slots and the output power actually went way down. Is this due to parallel capacitance adding up to the point that it is no longer tuned well for 10 meters? If I want to get to 3 or 4 watts with it is my best option to add an external amp? If necessary Ive got several old CB radios that I can hack and drive the driver amp with a single BS170. I¡¯ve also got a few IRF510 and IRF511 around, the latter which is out of production.
Thanks,
Jerry AC5JM
On Sep 22, 2023, at 8:02 AM, wb2cba@... wrote:
?
Hans,
As one of the 0.1% thank you for the encouragement to experiment on QDX. I was fearing to upset you with my QDX PA mod though it¡¯s just an innocent alternative PA test on QDX just to see what will be the outcome. Then I shared my experimentation with community as I always do.
As a fan of BS170 and using it in all my own rigs I can¡¯t agree more of your thoughts on BS170. You are absolutely right.
73
Barb WB2CBA
|
On 22/09/2023 17:38, John Z wrote: I think the total capacitance may
be too much for the chip to swing at 28 MHz. I think the reason is that they are not driven hard enough just by the Si5351. Maybe add a driver? 73 Alan GZFQ
|
Alan,
Same idea, expressed differently! A buffer/driver would definitely help!
JZ
On 22/09/2023 17:38, John Z wrote:
>? ?I think the total capacitance may
> be too much for the chip to swing at 28 MHz.
I think the reason is that they are not driven hard enough just by the
Si5351. Maybe add a driver?
73 Alan GZFQ
|
U3S user here for many years.
No point at all really using more than one BS170.
At 28Mhz using 12.0 volts and adjusting the Gate bias tune for 600mw max and that's it.
The lower bands will then produce upto 1.4 watts, 20m about 800mw.
But that assumes you keep you're antenna properly matched at ALL times.
These days I lower the volts to about 9.5v, gives 350mw at 10m, and remains quite flat across
all the bands, and ensures lots of headroom for when you have really bad loads, tested.
300mw is enough using a simple EFHW at 5m agl, just have a look at my wspr spots on the
database.
The South Pole, the odd VK, plenty of US stuff and usually into Brazil.
Have fun.
--
- 73 de Andy -
|
That's why other QRP Labs designs like the QCX and QDX have a driver stage. It's pretty simple, using some CMOS gates, but it's enough to drive the BS170 transistors. The U3S is a QRPp transmitter (under 1W) so it doesn't bother with the driver.
On Fri, Sep 22, 2023 at 1:45?PM Alan G4ZFQ < alan4alan@...> wrote: On 22/09/2023 17:38, John Z wrote:
>? ?I think the total capacitance may
> be too much for the chip to swing at 28 MHz.
I think the reason is that they are not driven hard enough just by the
Si5351. Maybe add a driver?
73 Alan GZFQ
|
It sounds like the UL3S is best suited for weak signal digital modes like WSPR and that the QCX- mini would be better suited for a CW beacon and the QCX has some other amazing attributes like CAT control which opens up the possibility of remote control which is beneficial since my 10m beacon is over 100 miles from my home QTH. ?I¡¯ve experimented with the CAT and it works very well. ?I have a QCX mini that I built then accidentally destroyed the CPU out of shear clumsiness but I went ahead and ordered a replacement today. ?I think it put out a solid 3 watts before I destroyed it. ?But I might also experiment with the UL3S to increase power output by adding a driver or an outboard amp. ?
toggle quoted message
Show quoted text
On Sep 22, 2023, at 2:14 PM, Shirley Dulcey KE1L <mark@...> wrote:
? That's why other QRP Labs designs like the QCX and QDX have a driver stage. It's pretty simple, using some CMOS gates, but it's enough to drive the BS170 transistors. The U3S is a QRPp transmitter (under 1W) so it doesn't bother with the driver. On Fri, Sep 22, 2023 at 1:45?PM Alan G4ZFQ < alan4alan@...> wrote: On 22/09/2023 17:38, John Z wrote:
>? ?I think the total capacitance may
> be too much for the chip to swing at 28 MHz.
I think the reason is that they are not driven hard enough just by the
Si5351. Maybe add a driver?
73 Alan GZFQ
|
Ryuji Suzuki AB1WX
Barb WB2CBA
Ted 2E0THH
Shirley Dulcey KE1L