Author Topic: 555 from transistors?  (Read 2693 times)

0 Members and 1 Guest are viewing this topic.

Offline BeaminTopic starter

  • Super Contributor
  • ***
  • Posts: 1567
  • Country: us
  • If you think my Boobs are big you should see my ba
555 from transistors?
« on: March 14, 2018, 10:49:40 am »
I counted 24 transistors in the diagram and few resistors and a diode.

Are these special transistors or diodes? Seems like pretty simple and as long as I'm running it slow it should be useable with through hole parts or even a bread board?

I was going to try and make something more complex like a LM358(not sure the name it has three leads: heat sink trans. package: to92 maybe? and will regulate voltages) but it has funky transistors with 3,4 or 5 /collectors/ emitters that seem like you couldn't replicate with a standard 2n2222 or "regular" generic transistors parts kits come with.

Do they make any kits where you learn a simple IC by making it transistor counterpart?

So after I get this working I'm going to build a ARM acorn with 583,927 discrete transistors. Some one should seriously try that. Someone on you tube has built and entire bread board computer using logic chips. Its all bread boards and through hole and about 4'X4' across and emulates an 8080 or arm or something. That's on my list of things to make I was going to try and sell it as art. Maybe some steem punk person will pay 300,000$ for it!
Max characters: 300; characters remaining: 191
Images in your signature must be no greater than 500x25 pixels
 

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 6136
  • Country: ro
Re: 555 from transistors?
« Reply #1 on: March 14, 2018, 11:02:49 am »
It has been done many times before, from simple integrated circuits to full computers.  Most probably no one will pay for it.
Here is one example (there are many smaller ones, including relays only, tubes only, transistors only, diodes only, and so on):


Offline BravoV

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: 00
  • +++ ATH1
Re: 555 from transistors?
« Reply #2 on: March 14, 2018, 11:03:41 am »

Offline BeaminTopic starter

  • Super Contributor
  • ***
  • Posts: 1567
  • Country: us
  • If you think my Boobs are big you should see my ba
Re: 555 from transistors?
« Reply #3 on: March 14, 2018, 11:10:41 am »


-> https://www.evilmadscientist.com/2013/555-kit/


Cool. That's what I was thinking. I would add LEDs to that so you could see wht was happening as it cycles.
Max characters: 300; characters remaining: 191
Images in your signature must be no greater than 500x25 pixels
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16510
  • Country: us
  • DavidH
Re: 555 from transistors?
« Reply #4 on: March 15, 2018, 05:21:06 am »
Are these special transistors or diodes?

The big difference is that the transistors and diodes are matched because of their monolithic construction which is difficult to impossible to duplicate with discrete parts.  Adding emitter degeneration can make up for some of this but introduces other compromises and it will be very difficult to get the same precision as the original parts.

One major problem is that PNP transistors built on an NPN process have very high base-emitter breakdown voltage which parts take advantage of to provide a high differential input voltage range for differential pairs.  Duplicating this without these PNP transistors requires a different circuit and again compromises precision.

Quote
I was going to try and make something more complex like a LM358(not sure the name it has three leads: heat sink trans. package: to92 maybe? and will regulate voltages) but it has funky transistors with 3,4 or 5 /collectors/ emitters that seem like you couldn't replicate with a standard 2n2222 or "regular" generic transistors parts kits come with.

Multiple collector and emitter transistors can be duplicated with multiple discrete devices but the matching problem mentioned above prevents doing this easily.  Of greater concern is that most schematics are either simplified or just equivalent to the actual part and leave out important features.  For instance most schematics of the LM358/LM324 do not show the transconductance reduction in the input differential pair, current mirrors, and biasing.

Quote
Do they make any kits where you learn a simple IC by making it transistor counterpart?

These days they use SPICE but in the past, special analog ICs were made which had the various integrated matched transistors for prototyping.

So what you are proposing can be done with discrete parts but it requires more than duplicating the circuit as shown.
 

Offline LeoTech

  • Regular Contributor
  • *
  • Posts: 75
  • Country: dk
  • High School student with a passion for electronics
Re: 555 from transistors?
« Reply #5 on: March 15, 2018, 12:15:04 pm »
Are these special transistors or diodes?

The big difference is that the transistors and diodes are matched because of their monolithic construction which is difficult to impossible to duplicate with discrete parts.  Adding emitter degeneration can make up for some of this but introduces other compromises and it will be very difficult to get the same precision as the original parts.

One major problem is that PNP transistors built on an NPN process have very high base-emitter breakdown voltage which parts take advantage of to provide a high differential input voltage range for differential pairs.  Duplicating this without these PNP transistors requires a different circuit and again compromises precision.

Quote
I was going to try and make something more complex like a LM358(not sure the name it has three leads: heat sink trans. package: to92 maybe? and will regulate voltages) but it has funky transistors with 3,4 or 5 /collectors/ emitters that seem like you couldn't replicate with a standard 2n2222 or "regular" generic transistors parts kits come with.

Multiple collector and emitter transistors can be duplicated with multiple discrete devices but the matching problem mentioned above prevents doing this easily.  Of greater concern is that most schematics are either simplified or just equivalent to the actual part and leave out important features.  For instance most schematics of the LM358/LM324 do not show the transconductance reduction in the input differential pair, current mirrors, and biasing.

Quote
Do they make any kits where you learn a simple IC by making it transistor counterpart?

These days they use SPICE but in the past, special analog ICs were made which had the various integrated matched transistors for prototyping.

So what you are proposing can be done with discrete parts but it requires more than duplicating the circuit as shown.

Dave actually made a video of that particular kit:



The video is quite interesting, he not only builds and tests the kit, but also explains the internal workings of a 555 timer ic - definitely check it out!

Leo
High School student with a passion and interest in electronics, both analog and digital!
 
The following users thanked this post: Beamin

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16510
  • Country: us
  • DavidH
Re: 555 from transistors?
« Reply #6 on: March 15, 2018, 04:21:31 pm »
They also make a 741 kit which is more illuminating about the changes which are necessary.  They added diodes in series with the emitters of the input differential pair transistors to support a high differential input voltage because the PNP cascodes which would normally do this cannot when discrete transistors are used.  (1) The 741 did not use transconductance reduction like the 324/358 so they did not have to worry about that and this is reflected in its high 33 picofarad compensation capacitor just like the integrated 741.

In order to model the input precision, 2 pairs of transistors and 1 pair of diodes would need to be matched.  They specify an input offset voltage of 2.0mV typical which I doubt they even came close to but their listed common mode rejection ratio of 43 dB which is much lower than an integrated 741 is probably about right.  The lack of emitter degeneration on the various current mirrors might explain the poor common mode rejection but I am not sure.  I would have expected it to be much higher even in a discrete design.

I would be impressed if they made a 324 kit with just one operational amplifier.  (2) Of course the difficult to implement transconductance reduction is not required to make a 324 class of operational amplifier; it just makes for a must less expensive part in IC form because the integrated compensation capacitor is much smaller saving area.

(1) This is not strictly true but special PNP "chopper" transistors which have questionable pricing and availability would need to be used.  It is much more practical to add diodes.

(2) It is possible to piece together the full 324 schematic and some datasheets include it or most of it.
 

Offline lordvader88

  • Frequent Contributor
  • **
  • !
  • Posts: 935
  • Country: ca
Re: 555 from transistors?
« Reply #7 on: March 16, 2018, 05:12:34 pm »
Ones of these boring days I am going to try for the fun of it.
 

Offline BeaminTopic starter

  • Super Contributor
  • ***
  • Posts: 1567
  • Country: us
  • If you think my Boobs are big you should see my ba
Re: 555 from transistors?
« Reply #8 on: March 17, 2018, 05:48:55 pm »
There is a good youtube video of a guy who dissolves IC and shows how they work under a microscope. Common sense says a PNP would be just a NPN with the metals inversed but it's not like that at all. That's what I like about this; everything is always more complex then what it seems.
Max characters: 300; characters remaining: 191
Images in your signature must be no greater than 500x25 pixels
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf