[f] AMB on Facebook

Understanding Beta

beta24

Re: Understanding Beta

Postby amb » September 24th, 2019, 5:09 pm

In one of the β24 prototypes I built, I had a BJT voltage follower driver stage before the output MOSFETs. Upon testing I was not happy with the outcome. The addition of the driver stage added one more pole to the amp’s ultrasonic response and required additional compensation to become stable. The added compensation reduced the overall bandwidth of the amp.

It should be noted that a benefit of MOSFETs is the fact that they are voltage-to-current transconductance devices with a very high input impedance, so a typical driver stage isn’t necessary other than to deal with gate capacitance. When it’s all said and done, I chose to eliminate the driver stage and use dynamic cascoding to minimize the issue, which also brings other advantages (higher bandwidth, more linear behavior, thermal distribution, etc.).

I think the employment of dynamic cascoding to every stage in the amp is a practice that isn’t common, and remains a hallmark of the β22 and β24 designs.

Alan0354 likes this post.
User avatar
amb
Site Admin
 
Posts: 10377
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Understanding Beta

Postby Alan0354 » September 25th, 2019, 9:33 am

Thanks

I am not particular familiar with MOSFET OPS, I am mainly a BJT designer, my amps are all BJT OPS. I am sure there are things that I don't know about. I do plan to design a MOSFET OPS in my near future.

Yes, BJT frequency response is slower, it's hard to find good driver transistors. What BJT did you use for driver that you were not happy with the result? I use 2SC4883 and 2SA1930 that has fT of 100MHz and 200MHz resp. These are the fastest driver transistors I could find. 2SA1930 is hard to find at the time, I bought like 200 of them in bulk. I have no intention to sell amps or pcb, this will last my life time. Most of the others are 30MHz or lower. I do have to have 3 compensation network to stabilize the amp to drive high capacitive load. My amp is not as fast as yours, it's only 350MHz and 33V/uS slew rate at large signal. I use LTSPICE, but I found it's quite useless for simulating large signal circuit on stability, must be the model of the transistors are not that accurate. Even the THD simulation is not accurate at all from my experience.

Yes, your cascode design will make it harder to have more than one output pair, you need 6 transistors to make one output pair, this will add up really fast. Also, the input capacitance of Q37 and Q38 are loading the output of the OPS, even with one pair like your design, if input capacitance of the MOSFET is 2000pF, it's already has 4000pF loading the OPS. This really adds up fast. Like for 4 output pairs, you will have 16,000pF loading the output, it's a lot of capacitive load.

What is Q21 and Q22 for, looks to be constant current source. I am not familiar with this design for MOSFET, are these MOSFETs DEPLETION mode MOSFET instead of ENHANCE mode MOSFET?

Sorry to get into this deep, I am just very interested in looking at different designs.

Alan
Alan0354
 
Posts: 12
Likes: 2 posts
Liked in: 0 post
Joined: September 24th, 2019, 3:53 pm
Location: Sunnyvale
Country: United States (us)

Re: Understanding Beta

Postby amb » September 25th, 2019, 11:09 am

I used BD139 and BD140 complementary pair as the driver stage, which have Ft of 160MHz. I tend to shy away from using hard-to-find Japanese transistors in my projects because they are intended to be built by many people, and specifying such parts can be a headache. The exception here are the 2SK170 and 2SJ74 input JFETs as there are no suitable substitutes, so I have to stock them in AMB audio shop as pre-matched quads. Needless to say, this represents a fairly expensive inventory to me, but it's what I must do to support my builders. I don't want to do this kind of thing for too many parts.

Yes, the cascode design makes it harder to add more parallel pairs of output devices. But for the β24, one set of output devices is enough. The Ciss of the N- and P-channel devices I use are 1400pF and 1300pF, respectively. But due to dynamic cascoding, the voltage across the gate, drain and source junctions of the main output MOSFETs are kept constant, so there is no capacitive charging taking place as the signal is swinging, leading to a effective nulling of the gate capacitance. Not only this, due to the fact that the output MOSFETs operate in virtual "stasis" (*), their effective transfer function also become linearized, approaching that of an ideal transistor.

Q21/Q22 are depletion mode MOSFETs serving as constant current sources in the circuit, to bias the zeners D9 and D10. These zeners are the ones keeping the cascode MOSFETs' gates (and sources) swinging along with the output signal, which provides the dynamic cascoding. Simple and effective!

(*) I use the word "stasis" here because it's a great description of what happens in the circuit, but it's not related to the "Stasis" power amps by Nelson Pass under the Threshold brand (or the Nakamichi branded versions which you list in your profile).
User avatar
amb
Site Admin
 
Posts: 10377
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Understanding Beta

Postby Alan0354 » September 26th, 2019, 7:19 pm

You use single jFET? I hoping I can find dual PFET.

What do you mean by "Stasis"? I really never under the meaning of Stasis even for the Nelson Pass's Stasis amp. I look at his schematic, it's just has feedback from VAS to IPS and OPS is Complementary feedback pair power transistors. That is using a composite NPN/PNP to get a super beta NPN as emitter follower. I never like CFP as the frequency response is much lower than simple EF. Most of the time in design, Less is More.
Alan0354
 
Posts: 12
Likes: 2 posts
Liked in: 0 post
Joined: September 24th, 2019, 3:53 pm
Location: Sunnyvale
Country: United States (us)

Re: Understanding Beta

Postby amb » September 26th, 2019, 8:36 pm

Alan0354 wrote:You use single jFET? I hoping I can find dual PFET.

Dual JFETs are also hard to find now. Toshiba 2SK389/2SJ109 would be appropriate but they haven't been manufactured for a long time. I use carefully hand-matched single devices, which can be matched better than the dual devices.

What do you mean by "Stasis"? I really never under the meaning of Stasis even for the Nelson Pass's Stasis amp. I look at his schematic, it's just has feedback from VAS to IPS and OPS is Complementary feedback pair power transistors. That is using a composite NPN/PNP to get a super beta NPN as emitter follower. I never like CFP as the frequency response is much lower than simple EF. Most of the time in design, Less is More.

According to thefreedictionary.com, one of the definitions is:

Stasis - A condition of balance among various forces; motionlessness.

In the case of β24 , due to dynamic cascode design, the voltage at each of the main output MOSFETs' three pins (gate, drain, source) swing together up and down by the same amount with the signal, there is only a constant DC offset amongst them. Since there is no relative voltage change across any of the three pins, I coin that condition "stasis". As I said previously, this condition is ideal, because the device operates at a single point along its transfer curve, making it super linear. There is also no charging/discharging of the junction capacitances, effectively raising the bandwidth of the devices and making it possible to drive directly from the VAS without a driver stage.

I have no idea what "Stasis" means in the Nelson Pass Threshold amps. I'm sure it has nothing to do with what I describe here.
User avatar
amb
Site Admin
 
Posts: 10377
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Understanding Beta

Postby Alan0354 » September 27th, 2019, 9:53 am

Yeh, even I use BJT LTP, I use BC550 and BC560 for my complementary IPS LTP. The beta of the match pairs are too low and draw too much base current for my liking. I have to hand pick each pair. This is actually not that hard compare picking 9 matching power transistors. I have to build a fixture to put like 50 transistor on and power up at the same time to measure the Vbe to pick match batches of 9. On top, I have to take into consideration the location of the transistors on the heatsink to compensate the different temperature difference. eg. I expect the transistors at the two ends will be a little lower temperature on the heatsink, so I have to choose two that the Vbe is about 3mV lower than the ones in the middle. I buy at least 100 at a time to pick out the sets.

One think I never got it that clear, what is the effect of Cbe ( or Cgs of MOSFET) of BJT on the driver. Seems like the Stasis condition you described is only in ideal situation. In real life, you have dVgs/dId = dVgs/dVout, as Id = Vout/RL

Seems like the effect of Cgs loading effect on the driver is divided by dVgs/dVout. So say Vgs increase by 1V when the ouput swing from 0 to 30V, so dVgs/dVout = 1/30. So if Cgs= 3000pF, the equivalent loading capacitance on the driver is 3000pF/30 = 100pF. What do you think? I have not seen anyone one or any book that talk about this.
Alan0354
 
Posts: 12
Likes: 2 posts
Liked in: 0 post
Joined: September 24th, 2019, 3:53 pm
Location: Sunnyvale
Country: United States (us)

Re: Understanding Beta

Postby Alan0354 » September 27th, 2019, 1:04 pm

I have a question, I am really not familiar with how to choose jFET. You choose 2SK170 and 2SJ74, I bought some J111 and J174 for reason they are commonly available. J111 and J174 both have lower capacitance, gm of J111 seems to be compatible with 2SK170, J174 is lower. What parameters are you looking for in choosing jFETs?

I took one of my amp and changed the input BJT to J111, it definitely change the sound, very obvious. I am neutral, it's just different. But a friend of mine really like the one with J111.

Actually the same question applies to BJT, I use BC550 and BC560 mainly because that's what a lot of people use, cheap and commonly available. I notice the beta are very high, you can get over 400 easily. The fT is high enough and capacitance is low. There is nothing I can say I don't like. But I have no idea what make a transistor "audio" transistor!!!

Maybe you can give me some insight on this.
Alan0354
 
Posts: 12
Likes: 2 posts
Liked in: 0 post
Joined: September 24th, 2019, 3:53 pm
Location: Sunnyvale
Country: United States (us)

Understanding Beta

Postby amb » September 28th, 2019, 11:51 am

Here is a simulation that shows the dynamic cascode in action. The amp is swinging 10Vrms sine wave into an 8Ω load. I put voltage probes on each of the 3 pins of Q35, one of the main output MOSFETs. The red trace is the voltage on the drain pin, the green trace is on the gate pin, and the blue trace is on the source pin. As you can see, all three swing together, so the voltage across any of the pins is a constant DC offset throughout. Vgs is simply the distance between the green and blue traces, Vds is the distance between the red and blue traces, and Vdg is the distance between the red and green traces. In each of these cases it is constant.

Image

Image

So despite that the MOSFET is delivering current into the load, it is in virtual "stasis".

As for JFET selection, the 2SK170/2SJ74 have very low noise and high gm characteristics. This is why they are used almost commonly in differential pair input stages. I have not found any other currently-available JFETs that perform as well. For other small signal transistors in this amp I use BC550C/BC560C.
User avatar
amb
Site Admin
 
Posts: 10377
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Understanding Beta

Postby Alan0354 » September 28th, 2019, 10:37 pm

I don't see you have a load resistor, you are not drawing any current, there should be no change in Vgs. If you draw current, it should be different.
Last edited by Alan0354 on September 28th, 2019, 10:48 pm, edited 1 time in total.
Alan0354
 
Posts: 12
Likes: 2 posts
Liked in: 0 post
Joined: September 24th, 2019, 3:53 pm
Location: Sunnyvale
Country: United States (us)

Re: Understanding Beta

Postby amb » September 28th, 2019, 10:41 pm

It’s not in the cropped area of the schematic, but there is indeed an 8Ω load resistor at the output of the amp, and there is 1.25A of current through the load.
User avatar
amb
Site Admin
 
Posts: 10377
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

PreviousNext

Return to β24 fully-differential power amplifier

Who is online

Users browsing this forum: Bing [Bot] and 2 guests

Quick Links

AMB audio DIY resources
AMB audio shop

cron