Hi,
I have tried to search other brands than IRF's V-FET's for fairly complementary pairs but don't find much at all.
Would be glad to receive advices on V-FET pairs that could be used as output devices in an amplifier.
So far I know at least Toshiba 2SJ200/-201 and 2SK1529/-1530 but they are rather expensive.
For the moment there's no power spec so any advice is thankful!
Cheers Michael
I have tried to search other brands than IRF's V-FET's for fairly complementary pairs but don't find much at all.
Would be glad to receive advices on V-FET pairs that could be used as output devices in an amplifier.
So far I know at least Toshiba 2SJ200/-201 and 2SK1529/-1530 but they are rather expensive.
For the moment there's no power spec so any advice is thankful!
Cheers Michael
There doesn't seem to be much.
I don't know what power level you're looking for.
These 30v TO-220 from On-Semi look remotely complementary, not great mind you.
MTP50P03HDL-D.pdf
NTP75N03L09-D.pdf
I don't know what power level you're looking for.
These 30v TO-220 from On-Semi look remotely complementary, not great mind you.
MTP50P03HDL-D.pdf
NTP75N03L09-D.pdf
Ultima Thule said:
Toshiba has new great pair: 2SK3497/2SJ618.
Accuphase uses them. Even more expensive.
Fairchild suggests their FQA19N20C and FQA12P20 as replacements. A quick glance at the data sheets shows about as close a compliment as the IR devices (not great) Both are about $2 @ 25 pieces from Mouser.
The IRFP9140 is a closer compliment to the IRFP240. If your application allows, you may want to look at the suggested alternative FQA17P10.
I haven't tried any of these yet, though.
The IRFP9140 is a closer compliment to the IRFP240. If your application allows, you may want to look at the suggested alternative FQA17P10.
I haven't tried any of these yet, though.
Re: Re: Complementary V-FET's other than IRF's???
interesting, but being even expesiver dosen't sound good... I will look up their datasheets, always interesting study new components.
Thanks Bob,
that could be something!
Let's keep on posting guys, I hope this thread could as well develope to a list on fairly complementary V-FET's from other brands everyone could enjoy of!
Cheers Michael
Tere Estman,estman said:
interesting, but being even expesiver dosen't sound good... I will look up their datasheets, always interesting study new components.
BobEllis said:
Thanks Bob,
that could be something!
Let's keep on posting guys, I hope this thread could as well develope to a list on fairly complementary V-FET's from other brands everyone could enjoy of!
Cheers Michael
I'm not sure if this is relevant, but a post in another forum suggested that BUZ900 could be subbed for 2SK135 and BUZ905 for 2SJ50
FAIRCHILD
Terve Michael !
FC has SFH9240, which do not suffer transconductance anomalies in audio band like IR's IRFP9240 does. I have put them together with IRFP244.
PS
Mulla lojuu niita laatikossa ...
Terve Michael !
FC has SFH9240, which do not suffer transconductance anomalies in audio band like IR's IRFP9240 does. I have put them together with IRFP244.
PS
Mulla lojuu niita laatikossa ...
dangus said:
Yes that's correct. The BUZ parts are simply Magnatec's version of the Hitachi parts. Exicon also do a smiliar line with their ECF and ECX parts. However none of these are really interchangeable with IRF style vertical MOSFETs as they are lateral MOSFETs.
Fairchild does not recommend the SFH9240 for new designs, which tells me that it won't be around long. That's why I didn't mention it, although I have used a couple tubes and am very happy with them.
Old farts remember Siemens BUZ9's as popular subs for Hitachi numbers.
Oh dear, that must have been before Magnatec started.
Shame that Siemens didn't have a proper P-channel for their BUZ23.
Re: FAIRCHILD
Tere jälle estman,
YES!! This is the main point with my thread, getting rid of IRF's P-ch FET's, in fact we could also propose just P-ch as well that could work well with some of IRF's N-ch FET's!
Cheers Michael
ps. kas sa tahad müda mulle SFH9240 või viru valge? 😉
Kus kohalt eestis oled?
estman said:
Tere jälle estman,
YES!! This is the main point with my thread, getting rid of IRF's P-ch FET's, in fact we could also propose just P-ch as well that could work well with some of IRF's N-ch FET's!
Cheers Michael
ps. kas sa tahad müda mulle SFH9240 või viru valge? 😉
Kus kohalt eestis oled?
P-ch
Moro!
Exactly. Probably these semis are designed in the same lab.
PS
Pohjarannikolla.
Koskenkorva & ViruValge: K-ch & V-ch😀
Ultima Thule said:
Moro!
Exactly. Probably these semis are designed in the same lab.
PS
Pohjarannikolla.
Koskenkorva & ViruValge: K-ch & V-ch😀
LTSpice VDMOS model for FQA12P20
Hi all,
There was some discussion about using the Fairchild fqa12p20 as a substitute for the IRFP9240.
I've been looking at this part, but couldn't find a SPICE model for it. Also, it's been noted that the standard SPICE NMOS and PMOS models (all levels 1, 2 and 3) only model the gate-drain, drain-source and gate-source capacitances as constants. No variations in these parameters with their corresponding voltages are taken into account..
I found some IRF subcircuit-level models for the IRFP244 and IRFP9240. These models use diode capacitances to predict variations of Cds and Cgd with voltage. I was using them in a fairly complex sim with LTSpice, and ran into some nasty problems where the DC operating point would not converge. In asking for help from the LTSpice user's group, a poster named analogspiceman told me these models were problematic, causing DC convergence failures. He suggested using the LTSpice-specific VDMOS models.
I ended up creating VDMOS models for the Fairchild fqa12p20 and the International Rectifier IRFP244, which appears to be a closer complement to the fqa12p20 than the IRFP240 is. These VDMOS models model the capacitance variation of Cgd and Cds with voltage, but assume Cgs is constant. After creating these models and putting them in my sim, the DC convergence problems were fixed. I'm posting this to share these models with you, so that you can use them if you like.
Let me describe how I fit the model parameters to the data. A while back, somebody pointed out to me that the solver plug-in for the Excel spreadsheet functions as an optimizer. You can have it vary multiple parameters in order to minimize some quantity. If that quantity represents the sum of squares of deviations from model-predicted values and datasheet values, then the solver can accomplish a nonlinear least-squares curve-fitting technique. This is how I've made the models. I do this curve fitting in the following ways (see the LTSpice docs for descriptions of the parameters):
1) Adjust the Cgdmin, Cgdmax and "a" parameters to match the model data of Cgd vs. Vds to the datasheet.
2) Obtain Cgs from the datasheet at Vds = -50V.
3) Adjust the Cjo, m and Vj body diode capacitance parameters to match the model data of Cds vs. Vds to the datasheet.
4) Adjust the Is, n and Rb body diode DC parameters to match the model data of body-diode forward-bias current to its voltage.
5) Adjust KP, Vto and Rs to match the model data of Id vs. Vgs to the datasheet values at the reference Vds=-40V. Assume LAMBDA is zero for this procedure
6) After the initial KP is found, try an initial value of LAMBDA to match the variation of Id with Vds at some convenient Vgs. For each new LAMBDA guess, recompute KP so that Id vs Vgs at Vds=-40V does not change. Keep trying different LAMBDA/KP value pairs until the variation of Id with Vds matches the datasheet as closely as possible.
Some of these curve-fitting procedures were done with more than 20 points used on the curve to determine the error. I've attached a spreadsheet that shows the calculations, as well as graphs that compare the datasheet parameters with the values predicted by the model. If you want to try the Excel, solver, you'll first need to enable it in Tools, add-ins. If it does not show up on the list of add-ins, you'll need to install it from the CD, then enable it. You will then get a Tools, Solver menu choice.
I've also attached models for the fqp3n25 and fqp2p25 devices, which were extracted in the same way. These look like good substitutes for the IRF610 and IRF9610 TO-220 devices. The IRFP244 model attached uses the IRF data for KP, LAMBDA and Rs, but uses the junction capacitance and body diode parameters as specified above.
Note that VDMOS is LTSpice-specific. Also, I did not try to match the Rds(on) values for using the devices as switches. The only way I know of to do this is to add a non-zero Rd, but this seems like it's not the thing to do if the device is to be used as a linear amplifier. I may just be paranoid about this 🙂
Hi all,
There was some discussion about using the Fairchild fqa12p20 as a substitute for the IRFP9240.
I've been looking at this part, but couldn't find a SPICE model for it. Also, it's been noted that the standard SPICE NMOS and PMOS models (all levels 1, 2 and 3) only model the gate-drain, drain-source and gate-source capacitances as constants. No variations in these parameters with their corresponding voltages are taken into account..
I found some IRF subcircuit-level models for the IRFP244 and IRFP9240. These models use diode capacitances to predict variations of Cds and Cgd with voltage. I was using them in a fairly complex sim with LTSpice, and ran into some nasty problems where the DC operating point would not converge. In asking for help from the LTSpice user's group, a poster named analogspiceman told me these models were problematic, causing DC convergence failures. He suggested using the LTSpice-specific VDMOS models.
I ended up creating VDMOS models for the Fairchild fqa12p20 and the International Rectifier IRFP244, which appears to be a closer complement to the fqa12p20 than the IRFP240 is. These VDMOS models model the capacitance variation of Cgd and Cds with voltage, but assume Cgs is constant. After creating these models and putting them in my sim, the DC convergence problems were fixed. I'm posting this to share these models with you, so that you can use them if you like.
Let me describe how I fit the model parameters to the data. A while back, somebody pointed out to me that the solver plug-in for the Excel spreadsheet functions as an optimizer. You can have it vary multiple parameters in order to minimize some quantity. If that quantity represents the sum of squares of deviations from model-predicted values and datasheet values, then the solver can accomplish a nonlinear least-squares curve-fitting technique. This is how I've made the models. I do this curve fitting in the following ways (see the LTSpice docs for descriptions of the parameters):
1) Adjust the Cgdmin, Cgdmax and "a" parameters to match the model data of Cgd vs. Vds to the datasheet.
2) Obtain Cgs from the datasheet at Vds = -50V.
3) Adjust the Cjo, m and Vj body diode capacitance parameters to match the model data of Cds vs. Vds to the datasheet.
4) Adjust the Is, n and Rb body diode DC parameters to match the model data of body-diode forward-bias current to its voltage.
5) Adjust KP, Vto and Rs to match the model data of Id vs. Vgs to the datasheet values at the reference Vds=-40V. Assume LAMBDA is zero for this procedure
6) After the initial KP is found, try an initial value of LAMBDA to match the variation of Id with Vds at some convenient Vgs. For each new LAMBDA guess, recompute KP so that Id vs Vgs at Vds=-40V does not change. Keep trying different LAMBDA/KP value pairs until the variation of Id with Vds matches the datasheet as closely as possible.
Some of these curve-fitting procedures were done with more than 20 points used on the curve to determine the error. I've attached a spreadsheet that shows the calculations, as well as graphs that compare the datasheet parameters with the values predicted by the model. If you want to try the Excel, solver, you'll first need to enable it in Tools, add-ins. If it does not show up on the list of add-ins, you'll need to install it from the CD, then enable it. You will then get a Tools, Solver menu choice.
I've also attached models for the fqp3n25 and fqp2p25 devices, which were extracted in the same way. These look like good substitutes for the IRF610 and IRF9610 TO-220 devices. The IRFP244 model attached uses the IRF data for KP, LAMBDA and Rs, but uses the junction capacitance and body diode parameters as specified above.
Note that VDMOS is LTSpice-specific. Also, I did not try to match the Rds(on) values for using the devices as switches. The only way I know of to do this is to add a non-zero Rd, but this seems like it's not the thing to do if the device is to be used as a linear amplifier. I may just be paranoid about this 🙂
Andy,
thank you for your work!
It will be a good idea, to post your models also to the SwitcherCAD III yahoo group.
Regards
Heinz!
thank you for your work!
It will be a good idea, to post your models also to the SwitcherCAD III yahoo group.
Regards
Heinz!
- Status
- Not open for further replies.