Turn on thump is usually related to an asymmetry in the caps and how they are charged on a differential input amp. As this is not differential input - I wonder if it is simply the slow RC time constant of the 15k resistor and 2200uF input stage supply?
I got a chance to put it on My A/B setup. Virtually indistinguishable from the VSSA, which of course is a compliment. Big sound stage, clear highs and tight deep bass. Very nice. I didn't try it with a poor PSU so I can't tell if all the extra caps are doing anything. I suspect they might help if using an SMPS, though some SMPS would have trouble charging those big caps.
Anyway, nice little amp.
Anyway, nice little amp.
Terry,
Thanks for the AB comparo. So, given that is virtually indistinguishable from VSSA - is there any reason to use more expensive lateral FETs when vertical FETs seem to do the job?
Thanks for the AB comparo. So, given that is virtually indistinguishable from VSSA - is there any reason to use more expensive lateral FETs when vertical FETs seem to do the job?
The VSSA has much fewer parts and sounds as good.
Laterals are more linear.
Laterals are tougher.
Laterals have built in protection diodes.
Laterals don't need temp compensation.
Other than that? No.
Laterals are more linear.
Laterals are tougher.
Laterals have built in protection diodes.
Laterals don't need temp compensation.
Other than that? No.
Last edited:
I think IRFP240/9240 also has built in protection diodes - they are quite durable too from what I have seen. I haven't yet burned out either though so can't say for sure.
Being able to get the Vishay hexFETs (genuine) for under $1ea is pretty good reason in my book.
Being able to get the Vishay hexFETs (genuine) for under $1ea is pretty good reason in my book.
Show me the data where you see the built in diodes in verticals.
You are ruled by discount. When you factor in the added complexity of using verticals as well as the additional parts required and the need to attach the VBE to the heatsink there is little savings. Then factor in the total cost of building a completed amp with PSU, speaker protection, heatsinks and case. A couple of dollars for superior outputs means very little.
You are ruled by discount. When you factor in the added complexity of using verticals as well as the additional parts required and the need to attach the VBE to the heatsink there is little savings. Then factor in the total cost of building a completed amp with PSU, speaker protection, heatsinks and case. A couple of dollars for superior outputs means very little.
I would love to live in the USA .
Easy access to mouser . digikey ... without expensive taxes.
I would never buy AliExpress components with these stores in my country
Easy access to mouser . digikey ... without expensive taxes.
I would never buy AliExpress components with these stores in my country
Unless I am misreading the diagram of the MOSFET, there appears to be a protection diode in place:
http://www.vishay.com/docs/91210/91210.pdf
A BD139 Vbe multiplier and two resistors costs about $0.20 (or less). I agree the PCB is a bit more complicated - but there's no difference in price to make it so.
A lateral FET costs about $5ea from a trusted US source. The IRFP is about $2 from Digikey. About a $6 savings for two per amp.
Not much in the scheme of things like $60 PSU and a $300 case. But with that comparison - not much matters from amp to amp - even ones that use dual die Alfets.
For DIYers who have a single PSU and make lots of different amps to try an play - the cost is a small factor. Mostly from the quest for bang for the buck. Even you like the fact that you can make a PCB for $1 and derive some satisfaction from it. I admit it, I love good amp deals and low cost amps that sounds good.
Even at $5ea - I have about 5 sets of lateral FETs myself. Some I even paid $12ea. Now that's a big difference compared to $1 hexFET.
I don't want to argue if one is better than the other. The laterals have their place and their fans. I guess I just like the verticals because, as you say, I am driven by good deals.
Last edited:
The diode is the intrinsic body diode, not the ones Terry is talking about. Laterals (the Renesas ones at least) have integrated zener diodes from gate to source to protect the gate. Verticals generally do not include those.
I think both are good, actually, though I prefer the hexfets because they have:
1. Much higher transconductance (gm) which means the OPS has higher linearity, that is, throughput is close to 1.
2. With traction and smps technologies, the big companies have put a lot of work into hexfets and they offer up to 5Kw dissipation and that wonderful cap drive system, the gate, which can be driven with almost no drive at all. It also means you can build a 500W amp with 1Kw devices which never break down under SOAR issues that kill bipolars.
3. Very good thermal characteristics that permit huge reserves of power.
4. Very low Rdson, down to a few milliohms, excellent for driving difficult speakers.
5. VERY CHEAP!!
The laterals are good too, but their gm is poor, their Rdson is high, their biggest is only a double die that is rated to 250W only. OTOH, they have a 'soft' sound, a bit tubey, and they are very easy to compensation thermally, essentially, they don't need any! And unfortunately they are expensive......
Hugh
1. Much higher transconductance (gm) which means the OPS has higher linearity, that is, throughput is close to 1.
2. With traction and smps technologies, the big companies have put a lot of work into hexfets and they offer up to 5Kw dissipation and that wonderful cap drive system, the gate, which can be driven with almost no drive at all. It also means you can build a 500W amp with 1Kw devices which never break down under SOAR issues that kill bipolars.
3. Very good thermal characteristics that permit huge reserves of power.
4. Very low Rdson, down to a few milliohms, excellent for driving difficult speakers.
5. VERY CHEAP!!
The laterals are good too, but their gm is poor, their Rdson is high, their biggest is only a double die that is rated to 250W only. OTOH, they have a 'soft' sound, a bit tubey, and they are very easy to compensation thermally, essentially, they don't need any! And unfortunately they are expensive......
Hugh
Hi Hugh,
Thanks for the technical reasons why you like hexFETs. Which one has 5kW of dissipation? That's a crazy amount of heat! 😱
Perhaps it's the high gm and low Rdson that makes me think they have more powerful bass drive capability?
Thanks for the technical reasons why you like hexFETs. Which one has 5kW of dissipation? That's a crazy amount of heat! 😱
Perhaps it's the high gm and low Rdson that makes me think they have more powerful bass drive capability?
You guys might be talking about diodes in two different places.
All MOSFETs have an intrinsic diode from drain to source.
The laterals (at least the Hitachi/Renesas version) also includes back to back zeners from gate to source - a nice feature that helps make them "tough".
See the old datasheet snip.
mlloyd1
All MOSFETs have an intrinsic diode from drain to source.
The laterals (at least the Hitachi/Renesas version) also includes back to back zeners from gate to source - a nice feature that helps make them "tough".
See the old datasheet snip.
mlloyd1
Attachments
Last edited:
This has been explained a few times over a few different threads recently. Seems $4 clouds the issue.
X, congratulations with a cool little amp! 
Terry did a great testing job, as he always does
Right, HexFETs have got no gate-to-source Zeners embedded. However, in EF + source follower arrangement, it's difficult to imagine a situation, where the sources don't follow the gates, damaging the FETs (unless one of the drivers shorts, but in this case Zener will not help either). I often use no Zeners in this kind of arrangement - never had any issues.
Terry did a great testing job, as he always does
Right, HexFETs have got no gate-to-source Zeners embedded. However, in EF + source follower arrangement, it's difficult to imagine a situation, where the sources don't follow the gates, damaging the FETs (unless one of the drivers shorts, but in this case Zener will not help either). I often use no Zeners in this kind of arrangement - never had any issues.
Thanks Valery! 🙂
I am looking forward to building my own as soon as the PCB arrives. Terry did a great job testing and debugging. Andrewlebon was very good with catching the error on the Vbe multiplier layout. AndewT helped with some excellent suggestions on improving layout. Both Idefixes and Sonal did a great job on the layouts. Very eye catching - and effective designs. It's neat to see how quickly an idea can become reality with group think.
I am looking forward to building my own as soon as the PCB arrives. Terry did a great job testing and debugging. Andrewlebon was very good with catching the error on the Vbe multiplier layout. AndewT helped with some excellent suggestions on improving layout. Both Idefixes and Sonal did a great job on the layouts. Very eye catching - and effective designs. It's neat to see how quickly an idea can become reality with group think.
Hugh, I agree with everything above - just one more thing, we normally have to consider throughout the design process - input capacitance.
Laterals have got lower Ciss, plus it does not really depend on Vds, allowing simple direct drive from the VAS stage with no issues.
HexFETs' higher Ciss pretty much depends on Vds, so I always tend to drive them with some sort of EF driver stage with low output impedance.
So, this is not really an issue - it's just a specialty, we have to take care of.
All the other properties make the HexFETs (and the newer TrenchFETs) excellent output devices.
