Remembering CMOS...

[Swedish Chef]

[quote author="Svart"] still thinking Chef!

i'm trying to finish this little assignment for the boss. [/quote]

No sweat! I am about to take the prototype to a producer called Alan :wink:

[quote author="Svart"] push one: R1 on, R2 off
push two: R1 off, R2 on
push three: R1 on, R2 on
[/quote]
that's it exactly!

[quote author="Svart"]sorry i didn't have much time to study the schematic, but will later today.[/quote]

No rush! :grin:
I'll be back later...

Cheers

chef :thumb:




:guinness: [/quote]

 

[Svart]

hey, i'm aking a prototype circuit for my boss named Alan!

how ironic.. is he scottish too?

:green:

 

[Swedish Chef]

Nahh, I think this alan is from Lincolnshire :green:
He has a very nice studio though!!!!

chef

 

[bcarso]

I'd use a couple of N channel MOSFETs to drive the relays---since these are inverting with the sources grounded you don't need the two inverters now used to drive the PNP's. Some TO92 case logic level parts with ~5-10 ohm Ron should be fine, unless those are really big relays.

Although the body diodes on some MOSFETs will act like zeners, and will also protect against reverse voltage, your relays when deenergized will have their lower ends go high. The simplest protection is a regular diode across the coil in reverse---1N4002 would work fine..

Why the two schottkies in series to the ultracap? I would think a single 1N4002 would suffice.

 

[Svart]

ok had a chance to actually sit down and look at the circuit better.

actually upon looking at it, I think that the control portion of it is pretty much as good as it's going to get. the first schmitt trigger is a good debounce measure. the 4081 *could* be taken out and the switch set to pull high instead since the 4027 senses on the rising pulse. what little i found on the 40106 tells me that it will pull high and low so you could indeed drive a pair of MOSFETs with it, however i still think the mosfets could be driven with the 4027 just fine.

i looked for some through hole MOSFETs in some kind of dual package and unfortunately most are SMD. most of the good discrete through-hole parts(with body diode) are fairly large(to-220). I'd still recommend maybe starting to work with SMD parts since this circuit could very well be compacted to the size of less than 1/3rd of a credit card.. and most everything is eventually going to end up SMD anyway.

i can come up with a layout if you need.

:thumb:

 

[Svart]

I see that Bcarso also agrees with the MOSFETs. I have had many occasions with relays that demand the use of BOTH of the diodes mentioned, the body diode AND the coil diode!

I suggested schottkys on the memory V line due to the least amount of forward drop so that 5v is only dropped a few % but i see he's using 5.5v so a normal diode will work here normally. Usually most ICs will work at 4.5v-5.5v but i've come across a few that once they dip below around 4.8v they start to act strangely so I typically use schottkys here.

two schottkys might be a little much but a lot of gear does series the diodes for better cost effective protection.

another suggestion.. if the unit that this board is going in doesn't need a dedicated logic VCC, then i would tap the 12v rail and use a zener-follower reg setup. (or a 7805 if you aren't too concerned about the rail linearity)


just some ideas!

 

[bcarso]

2N7000 FET's should work fine: TO-92 (through-hole) package, cheap, multiple-sourced, Ron typ at 4.5Vgs is 1.8 ohms, Ron max 5.3 ohms. I am assuming these are reasonably small relays with coil power no more than a few hundred milliwatts, hence current <100mA. If the relays are great big clunkers then bigger FET's are in order.

4000 series CMOS is not optimal for 5V operation but will work, and it does have much lower leakage/quiescent current than HC and other lower voltage series logic. It is certainly capable of driving a FET gate since there is no d.c. load and speed is of no importance.

 

[Swedish Chef]

[quote author="bcarso"] I'd use a couple of N channel MOSFETs to drive the relays---since these are inverting with the sources grounded you don't need the two inverters now used to drive the PNP's. [/quote]
The reason I swaped to the PNPs was because as it stood originally the NPNs seemed to be draining the Ultracap like billio but when I changed to PNPs it was fine. I know this means that the polarity of the switching is wrong but I can work those kind of probs out once I have a working scheme...!

[quote author="bcarso"] Some TO92 case logic level parts with ~5-10 ohm Ron should be fine, unless those are really big relays. [/quote]
The relays are approx. 500mW so no biggie.

[quote author="bcarso"] The simplest protection is a regular diode across the coil in reverse---1N4002 would work fine.. [/quote]

Yep those are in there but I didn't draw them! D'oh!

[quote author="bcarso"]Why the two schottkies in series to the ultracap? I would think a single 1N4002 would suffice. [/quote]
Again this comes down to leaking Ultracap. One leaked, two didn't, but it may have been something else compounding it :sad:

 

[Swedish Chef]

[quote author="Svart"] ok had a chance to actually sit down and look at the circuit better.
actually upon looking at it, I think that the control portion of it is pretty much as good as it's going to get. the first schmitt trigger is a good debounce measure. the 4081 *could* be taken out and the switch set to pull high instead since the 4027 senses on the rising pulse. what little i found on the 40106 tells me that it will pull high and low so you could indeed drive a pair of MOSFETs with it, however i still think the mosfets could be driven with the 4027 just fine. [/quote]
thank you both so much for the continued support!
So I can dump the 4081 IF I swap the polarity of the switch? Obviously the reason for having the 4081 was so that accidental switch knockage would not alter their states when the unit is off, but you are suggesting trying just flipping the polarity of the i/p to the 1st schmitt...Hmmm, interesting! :shock:

[quote author="Svart"] i looked for some through hole MOSFETs...fairly large(to-220). I'd still recommend maybe starting to work with SMD parts since this circuit could very well be compacted to the size of less than 1/3rd of a credit card.. and most everything is eventually going to end up SMD anyway. [/quote]
I have to hand build at least the first 25 of these blooming things! Do you want me to go completely mad? :green: I don't say 'No' but I do say 'Please Lord there must be an easier way...' :grin:

[quote author="Svart"]i can come up with a layout if you need. [/quote]

You may yet live to regret that :twisted:

Chef :thumb:

 

[Swedish Chef]

[quote author="Svart"] I see that Bcarso also agrees with the MOSFETs. I have had many occasions with relays that demand the use of BOTH of the diodes mentioned, the body diode AND the coil diode! [/quote]
:shock:
[quote author="Svart"] another suggestion.. if the unit that this board is going in doesn't need a dedicated logic VCC, then i would tap the 12v rail and use a zener-follower reg setup. (or a 7805 if you aren't too concerned about the rail linearity) [/quote]
Yep! A good point and one that I have breadboarded too :grin: Err can you breadboad a point Steve...? :?
[quote author="Svart"]just some ideas! [/quote]
No not 'Just'. Never 'Just'!!!! :thumb:

Chef

 

[Swedish Chef]

[quote author="bcarso"] 2N7000 FET's should work fine: TO-92 (through-hole) package, cheap, multiple-sourced, Ron typ at 4.5Vgs is 1.8 ohms, Ron max 5.3 ohms. I am assuming these are reasonably small relays with coil power no more than a few hundred milliwatts, hence current <100mA. If the relays are great big clunkers then bigger FET's are in order. [/quote]

Yep relays are 500mW so should be fine...

[quote author="bcarso"] 4000 series CMOS is not optimal for 5V operation but will work, and it does have much lower leakage/quiescent current than HC and other lower voltage series logic. [/quote]
The low current is FUNDAMENTAL as far as the 1F cap is concerned!

[quote author="bcarso"] It is certainly capable of driving a FET gate since there is no d.c. load and speed is of no importance. [/quote]
So when the unit is powered down what is the leakage like throught the FETs? That is the demon that I'm trying to drive out!!! :?

Cheers so much!

And dagnabit I need another cup of tea after all of that typing :thumb: and it's a little early in the day but :sam: :guinness:

chef

 

[Svart]

hey chef!

FETs are voltage driven. their gates need a charge(similar to charging a cap)to turn the device on. Current is not used to drive the gate like it is used to drive a BJT's base(however you can argue that BJTs are voltage driven also..). MOSFETs can actually stay on quite some time by themselves even if you unhook the gate pin completely because the gate stays charged. i've seen large die mosfets conduct simply because of the static on your finger charged the gate of the MOSFET while sitting on the desk not even installed on a board yet.. so really you would benefit from the MOSFET.

there should be virtually 0 current leakage from the gate to either the source or drain of the MOSFET.

the 4081 is an AND gate, and you have one input tied high. the switch pulls the other input low(through the schmitt) causing the output of the gate to go low. the pullup 10k does just that, pulling the and gate's output high again and triggering the 4027. You *could* remove the 4081 and move the switch to pull high and put the 10k in a pulldown setup. this is because the 4027 acts upon a positive going pulse, but in reality, you could really still have the switch pulldown, the system would just act upon the "buttonUP" action instead of the "buttonDown".

I would still leave the 40106 between he button and the 4027 as this is a good debounce circuit. you definately want that due to the switch's contact bounce, which might trigger more than one action from the 4027!

 

[Swedish Chef]

Cool! :thumb:
I'll get some parts together and give it a blast. Actually I'll be in France for next week, but will let you know how it goes asap!
I'm not sure it the ButtonUP thing would be too weird but I'll bread it and see how it feels...
Thanks so much for all the help :thumb:
Unfortunately I'll have to leave trying it out in the unit until I get it back from Alan but I can certainly breadboard it.
If Alan likes it as much as the last producer that used it, I will be very happy! :green:

chef

 

[Svart]

awesome! i'll have a bit of time the next few days, i might throw down an SMD board if you want.

:guinness:

 

[Swedish Chef]

Your boss must love you...! ;-)
throw it down kid, but only if you have some time to kill...

chef :thumb: :sam: :guinness: :green: :grin:

 

[Svart]

well my boss hasn't fired me yet... :shock:

being here 3 years makes me the longest lasting employee by far :wink:

 

[Swedish Chef]

so you bring him an apple every day or what? :green:

chef

 

[Svart]

he's scottish.. he doesn't eat apples..

when he insults me, i insult him back!

he seems to like it.. :shock:

 

[bcarso]

Does he bathe and use deodorant? I remember one job I was urged to interview for (by a hungry recruiter) and was relieved that the chief engineer found me insufficiently experienced. His body odor was appalling.

I suspect this practice drove away more than one former employee.

 

[Svart]

hmm the BO isn't that bad.. but the BREATH.... :roll:

he drinks nothing but coffee. and that breath is enough to make your eyes water.. no joke. and he's a close talker too..

where's the PUKE smiley?


ah but he leaves me to do my work most of the time... just most of the time everything is *RUSH RUSH RUSH RUSH*...

so should i layout for two SOT-23 FETs or for a SO-8 dual FET?.. hmm.

hey Chef, how many of these circuits do you need on one board?