I have been using Kicad for designing some circuits but I am not able to run sims on them. Mostly because some/a large number of the transistors that I have are not in the library and while I can label them what I want and assign the correct package to them. I have no idea how to get the data into the part to run a spice sim. I'm totally not being lazy on this one at all and been reading and searching for parts libraries and repositories w the parts I need. If anyone can break it down for me in 3rd grade language rather then programming language I would really appreciate any help.
I wasn't aware that the Kicad schematic could be used to create a netlist to run a simulation. I have drawn a schematic in LTSpice to simulate and then redrawn the schematic in Kicad to create a PCB.
KiCad does have some linkage to ngspice, but due to limited model availability I still use PChis method
I use LTspice for all circuit simulation and use KiCad and an old version of Orcad to draw schematics for publication
Isn't LTspice for only linear products? I'm using Fairchilds ons and exicons and a few others mostly. Also using some of the other linears stuff. I don't mind dragging the circuit into another program. The library is what is just absolutely ******* killing me now.
There are other components for ltspice around on the net.
I recently found some valve components to use in ltspice.
http://www.diyaudio.com/forums/software-tools/247614-ltspice-triode-unknown-subcircuit-2.html
You can use a huge range of other people's components in LTSpice provided that they are formatted correctly for LTSpice. I have a complete set of tube libraries that I use in it. (See the thread Nigel references for how to do it.)
I also use AD, and some other op-amp models in LTSpice, you can add them as subcircuits to the model libraries or use the .include option at the schematic level.
A number of different raw transistor models are supported, you just have to find the appropriate models.
There is a lot of stuff on using LTSpice on the web, a quick search is worth while.
http://www.linear.com/solutions/1083
I also use AD, and some other op-amp models in LTSpice, you can add them as subcircuits to the model libraries or use the .include option at the schematic level.
A number of different raw transistor models are supported, you just have to find the appropriate models.
There is a lot of stuff on using LTSpice on the web, a quick search is worth while.
http://www.linear.com/solutions/1083
As other posters have stated other models can be used in LTSpice. I have used transistor models from NXP and also models from Bob Cordell among others.
Before relying on the model accuracy it's worth checking them in test circuits to see if the data sheet typical curves are replicated (approximately). Then adjusting the parameters to suit.
Before relying on the model accuracy it's worth checking them in test circuits to see if the data sheet typical curves are replicated (approximately). Then adjusting the parameters to suit.
I heard that the data in the white sheets were more important on older documents with some omitting info like the mje 350 and 340. Supposedly the new white sheets are more solid. Would you concur?
Thanks Nigel, I know you helped before and i was able to find the sim downloads per your instructions for pretty much everything but getting it to mesh in Kicad took years off my life, gave me an ulcer, a goiter and could have been a possible smashing of my laptop. So LT spice is as good as it gets you reckon for sketching and simulation and then dump it into Kicad for PCB new? i think i have something solid but i am afraid of breakdown voltages on the LTP from the LSK389 which has been protected by being cascaded with mpsa42. so i should be able to have something solid. I see a lot of people using mje350 and 450 in the vas and while they may make decent drivers for lateral mosfets capacitance and speed seems to be an issue to be an issue unless i am cascoding them them which i guess is kinda wasted full when KSA1381 and the complement could be faster and better and make a simpler circuit. I still may drive with the MJEs because i have so many of them available still...With the laterals the way they are, would y'all recommend the trusted MJEs as drivers or do you think some enhancement mode mosfet may be a better choice? from all accounts the old crt drivers seem like the cats pajamas but they are hard to find other then from bdent who's charging like 8 bucks usd a pop for them. Basically using the parts and the higher voltage parts in cascode, i am thinking i should be able to have the same IPS and vas with minimal differences regardless of the number of output devices and hopefully have the option to run it in A or AB and have a simple selector switch to toggle modes. Does this sound too far fetched or should i stick with AB? i need a few amps for my trip amped towers w active crossing - 70w max on the tweeters, 150w on the high mids and 500w on the low mids...yes i know its is a low of power but theres a method to my madness especially if i can have a quick change from ab to a. the tweeters are 107 db if i am not mistaken, the high mids together are 100 as well as the low mids. i am sure a lot will be dissipated as heat with all things running but again the idea is to hopefully set them up where they are able to run max power in AB or run an A bias as well. I would have the 500w in the speaker with a passive XO with an option to run the active XO and 2 additional amps if the needs arise. any suggestions pointers etc other then - i lost my mind?
I am not sure I understand the comment about the white sheets.
I am looking at using the NXP PBHV8115X and PBHV9115X and the models appear to have about twice as much DC current gain compared with the Data Sheet typical curves.
At some point forget the simulations and turn the soldering iron on. There's only so much the simulation can do.
i apologize. it was a typo - i mean that the white sheets of older components were not as complete as the new one like on the MJE it doesn't state certain parameters that are critical while the newer ones have more of the data listed.
Love the way you think. I just need to sim a few things to make sure that certain things are correctly done to protect the lsk and other components so i don't burn the house down or blow everything up. I am not really concerned with numbers as much as i am with something sounding good.
In your opinion, do you think it is possible to over build an area? you hear people talk about sag with LFETs and if I really wanted to beef up the IPS and Vas so it can be anywhere from a 50w mono block to 1000w mono block in a or ab with the only difference being the ops is this a realistic design goal? ideally i would want it to be able to run A and AB - i would also like the 1st 2 states to be the same regardless of output and voltage.
Thanks for the clarification. No need to apologise. Communication and language are always open to different interpretations.
As you say the On Semiconductor MJE340 data sheet doesn't seem to specify any capacitance values or mention current-gain-bandwidth product but the (presumably newer) MJE243G does.
I don't have an opinion on over building an area. I guess that using the same driver for widely different voltages and loads is possible. It just depends on the circuit.
The Marantz Esotec MA 5 was selectable between Class A and Class B but it must have switched power supply voltages because the power output was either 30 W in Class A or 120 W in Class B. Which seems to me the way to go if switching amplifier class is a requirement.
As you say the On Semiconductor MJE340 data sheet doesn't seem to specify any capacitance values or mention current-gain-bandwidth product but the (presumably newer) MJE243G does.
I don't have an opinion on over building an area. I guess that using the same driver for widely different voltages and loads is possible. It just depends on the circuit.
The Marantz Esotec MA 5 was selectable between Class A and Class B but it must have switched power supply voltages because the power output was either 30 W in Class A or 120 W in Class B. Which seems to me the way to go if switching amplifier class is a requirement.
Why would the power supply voltages need to change? Let's say I get 500w at 4ohm on 78v main rails and 1kva toroid - secondary psu is 85v w 100va and have a AB current of 200. Increasing the q- current to 1.6 if everything is beefy should have me running approx 78watts if I do the circuit correctly with a linear power supply no?
So you feel Ltspice is the way to go to build a schematic and sim it and then flip to kicad to route the pcb?
So you feel Ltspice is the way to go to build a schematic and sim it and then flip to kicad to route the pcb?
Sims are useful but not everything.
Good PCB and component layout is vital.
I sim-ed an audio mixer and then laid out the pcb only get 1VAC hum on the output with input grounded !
The higher current paths were mixed in with the audio ground and was modulating it.
I reworked the pcb taking care of grounding and that fixed it.
An expensive lesson.
I imagine - a friend of mine was telling me about an older console mixers mute button was right in the vas and being there limited the pres swing and limited the head room. A lot to think about for sure. Speaking of mixers in still an old analog guy and love my Rane mixer and feel it needs some mods to make it perfect. Are you familiar with Dj audio mixers? Ideally it's not as loud as I would like for starters and I imagine that is easily fixed w a few resistors and allowing faster and more swing. I have to replace some fader so figured when I had it open may as well go for the gusto
- Status
- Not open for further replies.