Inverters /buffers in parallel

[George Herold]

Can I run two or three CMOS inverters (74HC14’s) in parallel to get
more current?

I’m using them as a monitor output with 500 ohms in series and then
50 ohms to ground. (Driving a cable to a ‘scope) That's a bit more
than the 5mA spec.

Should I add a bit of series resistance before the input to each
inverter? or before tying the outputs together?

George H.

 

[Sjouke Burry]

George Herold wrote:

Can I run two or three CMOS inverters (74HC14’s) in parallel to get
more current?

I’m using them as a monitor output with 500 ohms in series and then
50 ohms to ground. (Driving a cable to a ‘scope) That's a bit more
than the 5mA spec.

Should I add a bit of series resistance before the input to each
inverter? or before tying the outputs together?

George H.
I have often done so, with no ill effects.

Up to 8 gates, ttl as well as cmos.
Making the input slower, would not be a good idea,
because the period in which the gates disagree, becomes longer.
Which also means that the gates should come from one single chip,
because they would have a better chance of being of equal quality.

 

[John Larkin]

On Tue, 30 Nov 2010 12:03:58 -0800 (PST), George Herold
<ggherold@gmail.com> wrote:

Can I run two or three CMOS inverters (74HC14’s) in parallel to get
more current?

Sure. We do 6 at a time sometimes. But HC14s are kinda wimpy; the more
modern TinyLogic parts can drive harder.

I’m using them as a monitor output with 500 ohms in series and then
50 ohms to ground. (Driving a cable to a ‘scope) That's a bit more
than the 5mA spec.

At those levels, a single HC14 should work fine. With a 5 volt supply,
figure around 40 ohms equivalent drive impedance, so cut your first
resistor to around 400 ohms to get a volt out.

Should I add a bit of series resistance before the input to each
inverter? or before tying the outputs together?

Nope, just strap them together, inputs and outputs.

John

 

[krw@att.bizzzzzzzzzzzz]

On Tue, 30 Nov 2010 21:39:19 +0100, Sjouke Burry
<burrynulnulfour@ppllaanneett.nnll> wrote:

George Herold wrote:
Can I run two or three CMOS inverters (74HC14’s) in parallel to get
more current?

I’m using them as a monitor output with 500 ohms in series and then
50 ohms to ground. (Driving a cable to a ‘scope) That's a bit more
than the 5mA spec.

Should I add a bit of series resistance before the input to each
inverter? or before tying the outputs together?

George H.
I have often done so, with no ill effects.
Up to 8 gates, ttl as well as cmos.
Making the input slower, would not be a good idea,
because the period in which the gates disagree, becomes longer.
Which also means that the gates should come from one single chip,
because they would have a better chance of being of equal quality.

If they're on a single chip watch the maximum ratings on the Vcc/Gnd leads.
There will be a chip maximum current rating that often limits the number that
can paralleled.

 

[John Fields]

On Tue, 30 Nov 2010 12:44:07 -0800, John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Tue, 30 Nov 2010 12:03:58 -0800 (PST), George Herold
ggherold@gmail.com> wrote:

Can I run two or three CMOS inverters (74HC14’s) in parallel to get
more current?


Sure. We do 6 at a time sometimes. But HC14s are kinda wimpy; the more
modern TinyLogic parts can drive harder.


I’m using them as a monitor output with 500 ohms in series and then
50 ohms to ground. (Driving a cable to a ‘scope) That's a bit more
than the 5mA spec.

At those levels, a single HC14 should work fine. With a 5 volt supply,
figure around 40 ohms equivalent drive impedance, so cut your first
resistor to around 400 ohms to get a volt out.
^ ^^^^

---
No.

Since:

E1 * R3
E2 = --------------,
R1 + R2 + R3

then:

.. 5V E1
.. |
.. [40R]R1
.. |
.. [400R]R2
.. |
.. +---->0.51V E2
.. |
.. [50R]R3
.. |
.. GND

Should I add a bit of series resistance before the input to each
inverter? or before tying the outputs together?

Nope, just strap them together, inputs and outputs.

---
To avoid mismatch problems he should match the impedance of the
paralleled inveters to that of the cable, and then, to scale the input
to the scope properly, use a voltage divider on the output of the
cable which matches the cable impedance and a ratio which scales the
output to the input of the scope.

Using your 40 ohms as an example, and wanting a 10:1 probe using 50
ohm coax:



GEN 2.5V 50 OHM LINE
..+-----------+ \
..|5V---[40R]-|--[10R]-O-----//------O-+
..+-----------+ |
.. [45R] 0.25V
.. | /
.. +--+-->[SCOPE]
.. |
.. [5R]
.. |
.. |
.. GND


---
JF

 

[George Herold]

On Nov 30, 3:44 pm, John Larkin
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:

On Tue, 30 Nov 2010 12:03:58 -0800 (PST), George Herold

ggher...@gmail.com> wrote:
Can I run two or three CMOS inverters (74HC14’s) in parallel to get
more current?

Sure. We do 6 at a time sometimes. But HC14s are kinda wimpy; the more
modern TinyLogic parts can drive harder.

 I’m using them as a monitor output with 500 ohms in series and then
50 ohms to ground.   (Driving a cable to a ‘scope)  That's a bit more
than the 5mA spec.

At those levels, a single HC14 should work fine. With a 5 volt supply,
figure around 40 ohms equivalent drive impedance, so cut your first
resistor to around 400 ohms to get a volt out.



Should I add a bit of series resistance before the input to each
inverter?  or before tying the outputs together?

Nope, just strap them together, inputs and outputs.

John

Hmm, I must be missing something. I'm driving something near 500 ohms
with 5 volts... that's 10mA. Don't I want at least two? (Though I
started with only one and it worked fine.)

George H.

Oh thanks for the responses guys...

OK now I'm confused. I was looking at the spec sheet and it lists
maximum current as +/- 25mA. Where did the 5mA number come from?
It's listed on the digikey web site but I don't see it in the spec
sheet...

 

[George Herold]

On Dec 1, 8:35 am, John Fields <jfie...@austininstruments.com> wrote:

On Tue, 30 Nov 2010 12:44:07 -0800, John Larkin





jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
On Tue, 30 Nov 2010 12:03:58 -0800 (PST), George Herold
ggher...@gmail.com> wrote:

Can I run two or three CMOS inverters (74HC14 s) in parallel to get
more current?

Sure. We do 6 at a time sometimes. But HC14s are kinda wimpy; the more
modern TinyLogic parts can drive harder.

 I m using them as a monitor output with 500 ohms in series and then
50 ohms to ground.   (Driving a cable to a scope)  That's a bit more
than the 5mA spec.

At those levels, a single HC14 should work fine. With a 5 volt supply,
figure around 40 ohms equivalent drive impedance, so cut your first
resistor to around 400 ohms to get a volt out.

                                    ^ ^^^^
---
No.

Since:

             E1 * R3
     E2 = --------------,
           R1 + R2 + R3

then:

.    5V E1
.     |
.   [40R]R1
.     |
.   [400R]R2
.     |
.     +---->0.51V E2
.     |
.   [50R]R3
.     |
.    GND

Should I add a bit of series resistance before the input to each
inverter?  or before tying the outputs together?

Nope, just strap them together, inputs and outputs.

---
To avoid mismatch problems he should match the impedance of the
paralleled inveters to that of the cable, and then, to scale the input
to the scope properly, use a voltage divider on the output of the
cable which matches the cable impedance and a ratio which scales the
output to the input of the scope.

Using your 40 ohms as an example, and wanting a 10:1 probe using 50
ohm coax:

      GEN        2.5V   50 OHM LINE
.+-----------+       \
.|5V---[40R]-|--[10R]-O-----//------O-+
.+-----------+                        |
.                                   [45R]  0.25V
.                                     |   /  
.                                     +--+-->[SCOPE]
.                                     |
.                                    [5R]
.                                     |
.                                     |          
.                                    GND

---
JF- Hide quoted text -

- Show quoted text -

Your first picture is what I'm using. It works fine for driving a
coax to a 'scope. (something like a 3ns rise time and no ringy
dingies.)

 

[John Larkin]

On Wed, 01 Dec 2010 07:35:12 -0600, John Fields
<jfields@austininstruments.com> wrote:

On Tue, 30 Nov 2010 12:44:07 -0800, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Tue, 30 Nov 2010 12:03:58 -0800 (PST), George Herold
ggherold@gmail.com> wrote:

Can I run two or three CMOS inverters (74HC14’s) in parallel to get
more current?


Sure. We do 6 at a time sometimes. But HC14s are kinda wimpy; the more
modern TinyLogic parts can drive harder.


I’m using them as a monitor output with 500 ohms in series and then
50 ohms to ground. (Driving a cable to a ‘scope) That's a bit more
than the 5mA spec.

At those levels, a single HC14 should work fine. With a 5 volt supply,
figure around 40 ohms equivalent drive impedance, so cut your first
resistor to around 400 ohms to get a volt out.
^ ^^^^
---
No.

Since:

E1 * R3
E2 = --------------,
R1 + R2 + R3

then:

. 5V E1
. |
. [40R]R1
. |
. [400R]R2
. |
. +---->0.51V E2
. |
. [50R]R3
. |
. GND


Should I add a bit of series resistance before the input to each
inverter? or before tying the outputs together?

Nope, just strap them together, inputs and outputs.

---
To avoid mismatch problems he should match the impedance of the
paralleled inveters to that of the cable, and then, to scale the input
to the scope properly, use a voltage divider on the output of the
cable which matches the cable impedance and a ratio which scales the
output to the input of the scope.

Using your 40 ohms as an example, and wanting a 10:1 probe using 50
ohm coax:



GEN 2.5V 50 OHM LINE
.+-----------+ \
.|5V---[40R]-|--[10R]-O-----//------O-+
.+-----------+ |
. [45R] 0.25V
. | /
. +--+-->[SCOPE]
. |
. [5R]
. |
. |
. GND


---
JF

No. The HC gate won't look like 40 ohms if you load it that hard; it
is a mosfet after all, not a resistor. And the 40 ohm value is only
approximate anyhow.

It's easier and more precise to divide before the coax.

John

 

[John Larkin]

On Wed, 1 Dec 2010 08:50:55 -0800 (PST), George Herold
<ggherold@gmail.com> wrote:

On Nov 30, 3:44 pm, John Larkin
jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
On Tue, 30 Nov 2010 12:03:58 -0800 (PST), George Herold

ggher...@gmail.com> wrote:
Can I run two or three CMOS inverters (74HC14’s) in parallel to get
more current?

Sure. We do 6 at a time sometimes. But HC14s are kinda wimpy; the more
modern TinyLogic parts can drive harder.

 I’m using them as a monitor output with 500 ohms in series and then
50 ohms to ground.   (Driving a cable to a ‘scope)  That's a bit more
than the 5mA spec.

At those levels, a single HC14 should work fine. With a 5 volt supply,
figure around 40 ohms equivalent drive impedance, so cut your first
resistor to around 400 ohms to get a volt out.



Should I add a bit of series resistance before the input to each
inverter?  or before tying the outputs together?

Nope, just strap them together, inputs and outputs.

John

Hmm, I must be missing something. I'm driving something near 500 ohms
with 5 volts... that's 10mA. Don't I want at least two? (Though I
started with only one and it worked fine.)

George H.

Oh thanks for the responses guys...

OK now I'm confused. I was looking at the spec sheet and it lists
maximum current as +/- 25mA. Where did the 5mA number come from?
It's listed on the digikey web site but I don't see it in the spec
sheet...

If you look at the actual HC fet drain curves (in the databooks, but
not on individual datasheets), at Vcc=5, the initial e/i slope is
around 30-40 ohms up to a volt or so, which would correspond to a 4
volt pullup at 25-35 mA roughly. They are still sorta ohmic there. But
sure, use a couple in parallel... they're cheap enough.

John

 

[John Fields]

On Wed, 01 Dec 2010 21:09:36 -0800, John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Wed, 01 Dec 2010 07:35:12 -0600, John Fields
jfields@austininstruments.com> wrote:

On Tue, 30 Nov 2010 12:44:07 -0800, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Tue, 30 Nov 2010 12:03:58 -0800 (PST), George Herold
ggherold@gmail.com> wrote:

Can I run two or three CMOS inverters (74HC14’s) in parallel to get
more current?


Sure. We do 6 at a time sometimes. But HC14s are kinda wimpy; the more
modern TinyLogic parts can drive harder.


I’m using them as a monitor output with 500 ohms in series and then
50 ohms to ground. (Driving a cable to a ‘scope) That's a bit more
than the 5mA spec.

At those levels, a single HC14 should work fine. With a 5 volt supply,
figure around 40 ohms equivalent drive impedance, so cut your first
resistor to around 400 ohms to get a volt out.
^ ^^^^
---
No.

Since:

E1 * R3
E2 = --------------,
R1 + R2 + R3

then:

. 5V E1
. |
. [40R]R1
. |
. [400R]R2
. |
. +---->0.51V E2
. |
. [50R]R3
. |
. GND


Should I add a bit of series resistance before the input to each
inverter? or before tying the outputs together?

Nope, just strap them together, inputs and outputs.

---
To avoid mismatch problems he should match the impedance of the
paralleled inveters to that of the cable, and then, to scale the input
to the scope properly, use a voltage divider on the output of the
cable which matches the cable impedance and a ratio which scales the
output to the input of the scope.

Using your 40 ohms as an example, and wanting a 10:1 probe using 50
ohm coax:



GEN 2.5V 50 OHM LINE
.+-----------+ \
.|5V---[40R]-|--[10R]-O-----//------O-+
.+-----------+ |
. [45R] 0.25V
. | /
. +--+-->[SCOPE]
. |
. [5R]
. |
. |
. GND


---
JF


No. The HC gate won't look like 40 ohms if you load it that hard; it
is a mosfet after all, not a resistor. And the 40 ohm value is only
approximate anyhow.

It's easier and more precise to divide before the coax.

---
If "precise" is what matters, then your: "around 40 ohms" and your:
"cut your first resistor to around 400 ohms" are certainly out of
place.

As a matter of fact, there's no good reason why the 440 ohm resistor
should have been substituted for the 500 ohm one, except perhaps that
you thought a 10:1 divider would yield a 5:1 voltage difference
between input and output voltages.

In any case, since reflections don't seem to be a problem (probably
because of back-terminating the cable) and if precision _is_
important, I'd suggest that George use a single-turn carbon or cermet
pot like this:


.. 5V
.. |
.. [40R]
.. |
.. +---E1
.. |
.. [1k]<-+
.. | |
.. +----+ E2
.. | /
.. +---O------//-----O--[SCOPE]
.. |
.. [50R]
.. |
.. GND

and adjust it so that E2 = 0.1 E1, which will give him a calibrated
10:1 attenuator while conveniently allowing the issue of the gate's
output resistance to be moot.

BTW, here's a plot of typical resistance VS current for Philips' 74HC
standard outputs:

news:c82ff6989gsf3tirpv59lqcqhhj6d9qvdm@4ax.com

---
JF

 

[John Fields]

On Thu, 2 Dec 2010 07:08:28 -0800 (PST), George Herold
<ggherold@gmail.com> wrote:

Thanks John, I don't need the precission at all... (who really cares
if the pulse is 5 V or 4.5 volts.) I just wanted a monitor to see
when things are triggering.

---
Then your original divider with a a single gate driving it will be
fine.

---
JF

 

[George Herold]

On Dec 2, 7:06 am, John Fields <jfie...@austininstruments.com> wrote:

On Wed, 01 Dec 2010 21:09:36 -0800, John Larkin





jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
On Wed, 01 Dec 2010 07:35:12 -0600, John Fields
jfie...@austininstruments.com> wrote:

On Tue, 30 Nov 2010 12:44:07 -0800, John Larkin
jjlar...@highNOTlandTHIStechnologyPART.com> wrote:

On Tue, 30 Nov 2010 12:03:58 -0800 (PST), George Herold
ggher...@gmail.com> wrote:

Can I run two or three CMOS inverters (74HC14 s) in parallel to get
more current?

Sure. We do 6 at a time sometimes. But HC14s are kinda wimpy; the more
modern TinyLogic parts can drive harder.

 I m using them as a monitor output with 500 ohms in series and then
50 ohms to ground.   (Driving a cable to a scope)  That's a bit more
than the 5mA spec.

At those levels, a single HC14 should work fine. With a 5 volt supply,
figure around 40 ohms equivalent drive impedance, so cut your first
resistor to around 400 ohms to get a volt out.
                                   ^ ^^^^
---
No.

Since:

            E1 * R3
    E2 = --------------,
          R1 + R2 + R3

then:

.    5V E1
.     |
.   [40R]R1
.     |
.   [400R]R2
.     |
.     +---->0.51V E2
.     |
.   [50R]R3
.     |
.    GND

Should I add a bit of series resistance before the input to each
inverter?  or before tying the outputs together?

Nope, just strap them together, inputs and outputs.

---
To avoid mismatch problems he should match the impedance of the
paralleled inveters to that of the cable, and then, to scale the input
to the scope properly, use a voltage divider on the output of the
cable which matches the cable impedance and a ratio which scales the
output to the input of the scope.

Using your 40 ohms as an example, and wanting a 10:1 probe using 50
ohm coax:

     GEN        2.5V   50 OHM LINE
.+-----------+       \
.|5V---[40R]-|--[10R]-O-----//------O-+
.+-----------+                        |
.                                   [45R]  0.25V
.                                     |   /  
.                                     +--+-->[SCOPE]
.                                     |
.                                    [5R]
.                                     |
.                                     |          
.                                    GND

---
JF

No. The HC gate won't look like 40 ohms if you load it that hard; it
is a mosfet after all, not a resistor. And the 40 ohm value is only
approximate anyhow.

It's easier and more precise to divide before the coax.

---
If "precise" is what matters, then your: "around 40 ohms" and your:
"cut your first resistor to around 400 ohms"  are certainly out of
place.

As a matter of fact, there's no good reason why the 440 ohm resistor
should have been substituted for the 500 ohm one, except perhaps that
you thought a 10:1 divider would yield a 5:1 voltage difference
between input and output voltages.

In any case, since reflections don't seem to be a problem (probably
because of back-terminating the cable) and if precision _is_
important, I'd suggest that George use a single-turn carbon or cermet
pot like this:

.    5V  
.     |
.   [40R]
.     |
.     +---E1  
.     |
.    [1k]<-+
.     |    |
.     +----+              E2
.     |                  /
.     +---O------//-----O--[SCOPE]
.     |
.   [50R]
.     |
.    GND

and adjust it so that E2 = 0.1 E1, which will give him a calibrated
10:1 attenuator while conveniently allowing the issue of the gate's
output resistance to be moot.

BTW, here's a plot of typical resistance VS current for Philips' 74HC
standard outputs:

news:c82ff6989gsf3tirpv59lqcqhhj6d9qvdm@4ax.com

---
JF- Hide quoted text -

- Show quoted text -

Thanks John, I don't need the precission at all... (who really cares
if the pulse is 5 V or 4.5 volts.) I just wanted a monitor to see
when things are triggering.

George H.