Cascode input and VAS amplifier design, will it work :)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

12 years ago I build an bipolar version of the Elektor Electronics Cresendo amplifier, and it sounded great. This design used a Cascode VAS, which I think was the main reason for good THD performance for this amp.

Now I have made a new design, which is based on the Cresendo input stage, and Cascode VAS. The new thing I want to try is to use Cascode in the input stage, together with active current mirrors.

Im thinking of using Analog Devices SSM2220 PNP and SSM2212 NPN match transistors for the input stage and the Cascode input stage.

Has anyone experience with the combination of Cascode input stage + current mirrors ?

I have attached a preliminary schematic of my design.
(it may have some errors, as I have not build it yet ;) )

Any feedback is appreaciated ! :)

Regards
Flemming Bach
Denmark
 

Attachments

  • Design1 (Schematic).zip
    430.1 KB · Views: 512
A problem...

I believe you won't get a predictable quiescent current in the Q27-Q28-Q29-Q30 stage. You will have two dueling very high gain circuits, for the pull down and pull up...this is a version of the same issue that faces the Randy Slone circuit discussed at length elsewhere on the forum...I fear this is kind of a fatal flaw, without some surgery...
 
Member
Joined 2009
Paid Member
Bob Cordell in his recent book, Designing Audio Power Amplifiers, descibes a stabilized current mirror - see the attached schematic - which should work with a complementary differential input stage.
current mirror.JPG
Also, a cascode VAS has a high output impedance, so you need a high input impedance output stage. The one shown in your schematic, a simple emitter follower, would overload your VAS and degrade its performance. If you want to stick with BJTs, you'll probably need a triple emitter follower to keep VAS performance high.
 
Last edited:
Because LTP input diferential amp is voltage orientated circuit with high output impedance, it is natural to load it with high impedance interlink amplifying stage in "keep it simple" manner, like mosfet transistor or similar. In this way of solution you have nice transfer from pure voltage to pure current. ;)
 

Attachments

  • P-CAD EDA - [VACA2000].JPG
    P-CAD EDA - [VACA2000].JPG
    527.1 KB · Views: 937
Hi,

Thanks for feedback!

I was also suspicious of the DC behavior of the current mirror in the LTP.

Looking at Bob Cordell´s circuit in figure 8, on page 9:
http://www.cordellaudio.com/papers/MOSFET_Power_Amp.pdf

I have made a new circuit based on Bob´s design.
I think Bob´s explain how it works very well, so I wont repeat it here.

Any comments about using this concept ?

Is the Diodes (D2 and D3) in Bob´s design a good thing to have ?

To Alexcp: Yes you are right, my circuits needs some extra emitter follower so the VAS is not loaded.

BR
Flemming Bach
 

Attachments

  • NewCircuit1.jpg
    NewCircuit1.jpg
    156.9 KB · Views: 907
Hi,

I found that the Technics SEA3 amplifier, does use a current mirror in the LTP inputstage, which also happens to be connected to a Cascode couplede to the JFET input stage. I dont see that they do anything special to bias/balance the upper current mirror. But again they also do not have any source resistors in the JFET input stage, and they use dual transistors which are properbly pretty equal matched.

@Workhorse: Any comments on the current mirror in the LTP, that you mentioned was fatal ?

But Robert Cordells circuit of how to bias the current mirror does seem to a smart trick :)

BR
Flemming
 

Attachments

  • Technics-SEA3.jpg
    Technics-SEA3.jpg
    424.5 KB · Views: 803
I found that the Technics SEA3 amplifier, does use a current mirror in the LTP inputstage, which also happens to be connected to a Cascode couplede to the JFET input stage. I dont see that they do anything special to bias/balance the upper current mirror.

The problem has nothing to do with the cascode, it is related to the complementary push-pull second stage which needs some sort of bias arrangement. In the schematic you show there is no complementary push-pull but rather a differential second stage (some call this push-pull as well, but its surely not complementary).

Samuel
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.