|
Just compared 2SC5551 to 2SC3357 in an LZ1AQ amp. There is almost no difference in IP2/IP3 and frequency response. I was going to post to say the 2SC3357 are available, but looking at Digikey there are only 9 in stock. It says one week factory lead time and new product, so I assume they are still in production.
--
===================================================================
Mike M
|
The 2SC3357 = NE85634 was a CATV RF Transistor manufactured by the japanese semiconductor giant RENESAS (merger of Mitsubishi & Hitachi & NEC) and later CEL (CA/USA - NEC partner). Production was discontinued in 2018. Unitronic (Taiwan)? too produced RF-transistors labeled as 2SC3357, obsolete too. The NE86534 (2S3357) were frequently used in Japanese Radio devices. Clifton Labs Jack Smith used it in his Norton Amplifiers.
?
The 2SC3357 offered since recently by Digikey are sourced from the the Hong-Kong based company EVVOSEMI. Althoug they carry the label 2SC3357, RF transistors from different manufacturers will very likely meet particular rated key values of parts offered as 2S3357, but they must not be identical in all parameters. It's worth to give'm a try.
?
regards, Fred
?
?
|
Thanks Fred. I was not aware of the history. The ones I have were sourced from DigiKey, so they are the EVVOSEMI parts. I also tested them in the Ikin Norton amp and they are equivalent to the 2SC5551 there as well. My hope was to find a suitable substitute that is still in production. The BFU590 seems to be one of the few still produced. I plan to test some in the future.
--
===================================================================
Mike M
|
The?2SC3357 has a much lower voltage rating then the 2SC5551
?
Everett
In a message dated 2/7/2025 11:07:17 AM Central Standard Time, groups@... writes:
?
Thanks Fred. I was not aware of the history. The ones I have were sourced from DigiKey, so they are the EVVOSEMI parts. I also tested them in the Ikin Norton amp and they are equivalent to the 2SC5551 there as well. My hope was to find a suitable substitute that is still in production. The BFU590 seems to be one of the few still produced. I plan to test some in the future.
--
===================================================================
Mike M
|
From the data sheets here are the absolute Max Voltages. As you can see the 2SC3357 has lower VCBO and VCEO:
?
????? 2SC5551??? 2SC3357
VCBO??? 40???????? 20
VCEO??? 30???????? 12
VEBO? ?? 2??? ? ??? 3
?
Measuring on an LZ1AQ amp with 2SC3357 operating at 12V, here are the max voltages over all four transistors:
?
???? Measured?? Margin
VCBO?? 5.5?????? 14.5
VCEO?? 6.2??????? 5.8
VEBO?? 0.743??? ? 2.257
?
?
So for the LZ1AQ design the 2SC3357 operates with plenty of margin.
--
===================================================================
Mike M
|
Mike,
It does make much difference anyway, as Digiky no longer have them.
toggle quoted message
Show quoted text
On Feb 7, 2025, at 3:37?PM, Mike M <groups@...> wrote:
? From the data sheets here are the absolute Max Voltages. As you can see the 2SC3357 has lower VCBO and VCEO:
?
????? 2SC5551??? 2SC3357
VCBO??? 40???????? 20
VCEO??? 30???????? 12
VEBO? ?? 2??? ? ??? 3
?
Measuring on an LZ1AQ amp with 2SC3357 operating at 12V, here are the max voltages over all four transistors:
?
???? Measured?? Margin
VCBO?? 5.5?????? 14.5
VCEO?? 6.2??????? 5.8
VEBO?? 0.743??? ? 2.257
?
?
So for the LZ1AQ design the 2SC3357 operates with plenty of margin.
--
===================================================================
Mike M
|
All of these BFU5xx series from NXP seem to have a good high ft.? However, if you consult the gain vs. frequency, all of them pretty much die out above 1.8 GHz.? NXP does not rate gain or noise figure above 1.8 GHz.? With an ft of 9 GHz on the BFU550, I'd expect FAR BETTER GAIN above 1.8 GHz, but it just isn't there.?
Yes, I realize ft is technically the maximum frequency of oscillation or where the silicon can overcome losses to oscillate.? But, why?? Here are some very telling plots from the NXP data sheet for the BFU550:
MSG = Maximum Stable gain. |s21|^2 = 10 log [s21]^2
Nothing useful above 1.8 GHz.? NXP rates the "Transition Frequency" measured at 25 mA and 900 MHz at 11 GHz.? OK.? This isn't what I learned long ago and far away in my EE classes.
Well, I looked it up online.? The transition frequency is the frequency at which the gain drops to unity.? Therefore that would be the same as the maximum frequency of oscillation.? So........how do they measure 11 GHz when operating at 900 MHz??? Seems the max frequency of oscillation would be far easier to obtain.
Dave - W?LEV
From the data sheets here are the absolute Max Voltages. As you can see the 2SC3357 has lower VCBO and VCEO:
?
????? 2SC5551??? 2SC3357
VCBO??? 40???????? 20
VCEO??? 30???????? 12
VEBO? ?? 2??? ? ??? 3
?
Measuring on an LZ1AQ amp with 2SC3357 operating at 12V, here are the max voltages over all four transistors:
?
???? Measured?? Margin
VCBO?? 5.5?????? 14.5
VCEO?? 6.2??????? 5.8
VEBO?? 0.743??? ? 2.257
?
?
So for the LZ1AQ design the 2SC3357 operates with plenty of margin.
--
===================================================================
Mike M
|
While you are in the S parameters (wish they would expressly state the reference impedance), the first numbers of the S parameter pairs is the magnitude. Second is the angle.
If you divide the |S21| by |S12| that is the "K" factor which must be larger than unity for stability. Less than unity is "unstable".
For example an NXP BUF520 shows a K of 6.8 at 6 ghz. at 8V/20 ma and around 4000 at 40 mhz.
Jim/VEZ
?
On Friday, February 7, 2025 at 05:16:32 PM EST, W0LEV via groups.io <davearea51a@...> wrote:
All of these BFU5xx series from NXP seem to have a good high ft.? However, if you consult the gain vs. frequency, all of them pretty much die out above 1.8 GHz.? NXP does not rate gain or noise figure above 1.8 GHz.? With an ft of 9 GHz on the BFU550, I'd expect FAR BETTER GAIN above 1.8 GHz, but it just isn't there.?
Yes, I realize ft is technically the maximum frequency of oscillation or where the silicon can overcome losses to oscillate.? But, why?? Here are some very telling plots from the NXP data sheet for the BFU550:
MSG = Maximum Stable gain. |s21|^2 = 10 log [s21]^2
Nothing useful above 1.8 GHz.? NXP rates the "Transition Frequency" measured at 25 mA and 900 MHz at 11 GHz.? OK.? This isn't what I learned long ago and far away in my EE classes.
Well, I looked it up online.? The transition frequency is the frequency at which the gain drops to unity.? Therefore that would be the same as the maximum frequency of oscillation.? So........how do they measure 11 GHz when operating at 900 MHz??? Seems the max frequency of oscillation would be far easier to obtain.
Dave - W?LEV
From the data sheets here are the absolute Max Voltages. As you can see the 2SC3357 has lower VCBO and VCEO:
?
????? 2SC5551??? 2SC3357
VCBO??? 40???????? 20
VCEO??? 30???????? 12
VEBO? ?? 2??? ? ??? 3
?
Measuring on an LZ1AQ amp with 2SC3357 operating at 12V, here are the max voltages over all four transistors:
?
???? Measured?? Margin
VCBO?? 5.5?????? 14.5
VCEO?? 6.2??????? 5.8
VEBO?? 0.743??? ? 2.257
?
?
So for the LZ1AQ design the 2SC3357 operates with plenty of margin.
--
===================================================================
Mike M
|
Not quite that simple.? Here is the complete eqn for calculating the Linvill Stability Factor, K, or in this presentation, C:
Dave - W?LEV
On Sat, Feb 8, 2025 at 5:08?PM James Redding WA9VEZ via <wa9vez= [email protected]> wrote:
While you are in the S parameters (wish they would expressly state the reference impedance), the first numbers of the S parameter pairs is the magnitude. Second is the angle.
If you divide the |S21| by |S12| that is the "K" factor which must be larger than unity for stability. Less than unity is "unstable".
For example an NXP BUF520 shows a K of 6.8 at 6 ghz. at 8V/20 ma and around 4000 at 40 mhz.
Jim/VEZ
?
On Friday, February 7, 2025 at 05:16:32 PM EST, W0LEV via <davearea51a= [email protected]> wrote:
All of these BFU5xx series from NXP seem to have a good high ft.? However, if you consult the gain vs. frequency, all of them pretty much die out above 1.8 GHz.? NXP does not rate gain or noise figure above 1.8 GHz.? With an ft of 9 GHz on the BFU550, I'd expect FAR BETTER GAIN above 1.8 GHz, but it just isn't there.?
Yes, I realize ft is technically the maximum frequency of oscillation or where the silicon can overcome losses to oscillate.? But, why?? Here are some very telling plots from the NXP data sheet for the BFU550:
MSG = Maximum Stable gain. |s21|^2 = 10 log [s21]^2
Nothing useful above 1.8 GHz.? NXP rates the "Transition Frequency" measured at 25 mA and 900 MHz at 11 GHz.? OK.? This isn't what I learned long ago and far away in my EE classes.
Well, I looked it up online.? The transition frequency is the frequency at which the gain drops to unity.? Therefore that would be the same as the maximum frequency of oscillation.? So........how do they measure 11 GHz when operating at 900 MHz??? Seems the max frequency of oscillation would be far easier to obtain.
Dave - W?LEV From the data sheets here are the absolute Max Voltages. As you can see the 2SC3357 has lower VCBO and VCEO:
?
????? 2SC5551??? 2SC3357
VCBO??? 40???????? 20
VCEO??? 30???????? 12
VEBO? ?? 2??? ? ??? 3
?
Measuring on an LZ1AQ amp with 2SC3357 operating at 12V, here are the max voltages over all four transistors:
?
???? Measured?? Margin
VCBO?? 5.5?????? 14.5
VCEO?? 6.2??????? 5.8
VEBO?? 0.743??? ? 2.257
?
?
So for the LZ1AQ design the 2SC3357 operates with plenty of margin.
--
===================================================================
Mike M
--
|
Ok......OK........I'll attach the complete expression for the Linvill Stability Factor.
Dave - W ?LEV
Not quite that simple.? Here is the complete eqn for calculating the Linvill Stability Factor, K, or in this presentation, C:
Dave - W?LEV
On Sat, Feb 8, 2025 at 5:08?PM James Redding WA9VEZ via <wa9vez= [email protected]> wrote:
While you are in the S parameters (wish they would expressly state the reference impedance), the first numbers of the S parameter pairs is the magnitude. Second is the angle.
If you divide the |S21| by |S12| that is the "K" factor which must be larger than unity for stability. Less than unity is "unstable".
For example an NXP BUF520 shows a K of 6.8 at 6 ghz. at 8V/20 ma and around 4000 at 40 mhz.
Jim/VEZ
?
On Friday, February 7, 2025 at 05:16:32 PM EST, W0LEV via <davearea51a= [email protected]> wrote:
All of these BFU5xx series from NXP seem to have a good high ft.? However, if you consult the gain vs. frequency, all of them pretty much die out above 1.8 GHz.? NXP does not rate gain or noise figure above 1.8 GHz.? With an ft of 9 GHz on the BFU550, I'd expect FAR BETTER GAIN above 1.8 GHz, but it just isn't there.?
Yes, I realize ft is technically the maximum frequency of oscillation or where the silicon can overcome losses to oscillate.? But, why?? Here are some very telling plots from the NXP data sheet for the BFU550:
MSG = Maximum Stable gain. |s21|^2 = 10 log [s21]^2
Nothing useful above 1.8 GHz.? NXP rates the "Transition Frequency" measured at 25 mA and 900 MHz at 11 GHz.? OK.? This isn't what I learned long ago and far away in my EE classes.
Well, I looked it up online.? The transition frequency is the frequency at which the gain drops to unity.? Therefore that would be the same as the maximum frequency of oscillation.? So........how do they measure 11 GHz when operating at 900 MHz??? Seems the max frequency of oscillation would be far easier to obtain.
Dave - W?LEV From the data sheets here are the absolute Max Voltages. As you can see the 2SC3357 has lower VCBO and VCEO:
?
????? 2SC5551??? 2SC3357
VCBO??? 40???????? 20
VCEO??? 30???????? 12
VEBO? ?? 2??? ? ??? 3
?
Measuring on an LZ1AQ amp with 2SC3357 operating at 12V, here are the max voltages over all four transistors:
?
???? Measured?? Margin
VCBO?? 5.5?????? 14.5
VCEO?? 6.2??????? 5.8
VEBO?? 0.743??? ? 2.257
?
?
So for the LZ1AQ design the 2SC3357 operates with plenty of margin.
--
===================================================================
Mike M
--
--
|
Let's not step into "non-Foster" antennas. Heads might explode.
Interesting section on loop antennas in Walter H Carlton's, Traveling Wave Antennas? on page 58.
Jim/VEZ
On Saturday, February 8, 2025 at 12:24:39 PM EST, W0LEV via groups.io <davearea51a@...> wrote:
Ok......OK........I'll attach the complete expression for the Linvill Stability Factor.
Dave - W ?LEV
Not quite that simple.? Here is the complete eqn for calculating the Linvill Stability Factor, K, or in this presentation, C:
Dave - W?LEV
On Sat, Feb 8, 2025 at 5:08?PM James Redding WA9VEZ via <wa9vez= [email protected]> wrote:
While you are in the S parameters (wish they would expressly state the reference impedance), the first numbers of the S parameter pairs is the magnitude. Second is the angle.
If you divide the |S21| by |S12| that is the "K" factor which must be larger than unity for stability. Less than unity is "unstable".
For example an NXP BUF520 shows a K of 6.8 at 6 ghz. at 8V/20 ma and around 4000 at 40 mhz.
Jim/VEZ
?
On Friday, February 7, 2025 at 05:16:32 PM EST, W0LEV via <davearea51a= [email protected]> wrote:
All of these BFU5xx series from NXP seem to have a good high ft.? However, if you consult the gain vs. frequency, all of them pretty much die out above 1.8 GHz.? NXP does not rate gain or noise figure above 1.8 GHz.? With an ft of 9 GHz on the BFU550, I'd expect FAR BETTER GAIN above 1.8 GHz, but it just isn't there.?
Yes, I realize ft is technically the maximum frequency of oscillation or where the silicon can overcome losses to oscillate.? But, why?? Here are some very telling plots from the NXP data sheet for the BFU550:
MSG = Maximum Stable gain. |s21|^2 = 10 log [s21]^2
Nothing useful above 1.8 GHz.? NXP rates the "Transition Frequency" measured at 25 mA and 900 MHz at 11 GHz.? OK.? This isn't what I learned long ago and far away in my EE classes.
Well, I looked it up online.? The transition frequency is the frequency at which the gain drops to unity.? Therefore that would be the same as the maximum frequency of oscillation.? So........how do they measure 11 GHz when operating at 900 MHz??? Seems the max frequency of oscillation would be far easier to obtain.
Dave - W?LEV From the data sheets here are the absolute Max Voltages. As you can see the 2SC3357 has lower VCBO and VCEO:
?
????? 2SC5551??? 2SC3357
VCBO??? 40???????? 20
VCEO??? 30???????? 12
VEBO? ?? 2??? ? ??? 3
?
Measuring on an LZ1AQ amp with 2SC3357 operating at 12V, here are the max voltages over all four transistors:
?
???? Measured?? Margin
VCBO?? 5.5?????? 14.5
VCEO?? 6.2??????? 5.8
VEBO?? 0.743??? ? 2.257
?
?
So for the LZ1AQ design the 2SC3357 operates with plenty of margin.
--
===================================================================
Mike M
--
--
|
On Thu, Feb 6, 2025 at 04:39 PM, Mike M wrote:
Just compared 2SC5551 to 2SC3357 in an LZ1AQ amp. There is almost no difference in IP2/IP3 and frequency response.
My simulation results mirror your experimental(??) findings. There is negligible IP3 & frequency response differences between 2SC5551 and 2SC3357.?
?
I used the circuit model that I previously developed for my prototype LZ1AQ and detailed in message #20829. The 2SC3357 model is from NEC, while the 2SC5551 model is from LTWiki. The latter is just the chip model & does NOT include the SOT-89 package parasitics. However, I suspect the package parasitics can be ignored at HF.?
?
In the IP3 comparison, 2SC3357 (blue) beats 2SC5551 (green) by ~1 dB, while 2N2222A (red) is 3~4 dB lower than the rest. I am surprised that the 2N2222A is so much poorer, because, in general, IP3 is a function of the supply current and the 3 BJTs have almost similar simulated Icc values.?
?
Frequency response: 2SC5551 (green) and 2SC3357 (blue) traces nearly overlapped. Again, the 2N2222A (red) is clearly inferior to the other two; ~4 dB lower gain at the 29 MHz peak.
?
73, Leong, 9M2LCL, ex-9W2LC.?
|
Thanks for replying. Always good when simulations match measurements! I have since tested the same circuit with BFU590QX, and they perform about the same. So we have three transistors that perform about the same and the 2N2222 that is not as good. My only test with 2N2222 was at 10V. I'll test at 12V soon, but I expect the results will not be as good as the other three.
--
===================================================================
Mike M
|
Just an update. DigiKey has 2SC3357 back in stock. They only have 1009, so it looks like they just got one reel in. They are listed as a "new" part, so I think the issue has been supply chain related not that the parts are obsolete. They are $0.27 in quantity one and $0.18 in quantity 10, so less expensive than alternatives like BFU590QX.
?
--
===================================================================
Mike M
|
On Sat, Mar 1, 2025 at 05:04 PM, Mike M wrote:
They are listed as a "new" part,
Yes, because it is not the genuine 2SC3357 from NEC/Unisonic, which is obsolete since a couple of years. It is a new part from EVVOSEMI, a fabless Hong-Kong based company, specialised in second sourcing of obsolete discrete semiconductors.
?
regards, Fred
|
Everett N4CY
James Redding WA9VEZ