Reply by Daniel Pitts March 12, 20142014-03-12
On 3/12/14 9:33 AM, Lasse Langwadt Christensen wrote:
> Den onsdag den 12. marts 2014 03.10.34 UTC+1 skrev Daniel Pitts: >> On 3/11/14 4:39 PM, Maynard A. Philbrook Jr. wrote: > > just remember that now that you have turned things up side down the Vgs is you supply - Vol of your controller > So if you supply your LEDs with 1.8V it'll be at most 1.8V
As I wrote earlier:
>newsgroup.nospam@virtualinfinity.net says... >> Looking more carefully, I think I might need P-Channel MOSFETS, and then >> I'd need to invert my logic driver (replace the 74HC238 with a 74HC138).
Now, my source is VCC(+5v) always, The Vgs will be 0v (off) when the gate voltage is VCC(+5v). The Vgs will be -5v (on) when my gate voltage is at GND. The 74HC138 will select exactly one of these, the rest will be near VCC. The one that is "logic-enabled" will be near GND. My schematic: <http://virtualinfinity.net/electronics/LED.sch.pdf> My board layout: <http://virtualinfinity.net/electronics/LED.brd.pdf>
Reply by Lasse Langwadt Christensen March 12, 20142014-03-12
Den onsdag den 12. marts 2014 03.10.34 UTC+1 skrev Daniel Pitts:
> On 3/11/14 4:39 PM, Maynard A. Philbrook Jr. wrote: >=20 > > In article <71JTu.1700$KZ7.1308@fx32.iad>, >=20 > > newsgroup.nospam@virtualinfinity.net says... >=20 > >> Looking more carefully, I think I might need P-Channel MOSFETS, and th=
en
>=20 > >> I'd need to invert my logic driver (replace the 74HC238 with a 74HC138=
).
>=20 > >> If I used an N-Channel, I don't think I could get the Vgs to be hig=
h
>=20 > >> enough. >=20 > >> >=20 > >> I was thinking this part: >=20 > >> http://www.digikey.com/product-detail/en/SCH1302-TL-E/869-1194-1-ND/21=
65672
>=20 > >> >=20 > >> Then I would have the source connected to +5v, the drain connected to =
to
>=20 > >> my load, and the gate connected to the output of the 1-of-8. >=20 > >> >=20 > >> The load is some parallel LEDs (each in series with a >=20 > >> constant-current-sink), so the drain voltage could vary between +1.8v =
up
>=20 > >> to +5.0v, The current can vary from 10mA to 1200mA. Depending on how >=20 > >> many parallel LEDs are enabled. >=20 > >> >=20 > >> Does this sound like it would work? >=20 > >> >=20 > >> Thanks, >=20 > >> Daniel. >=20 > >> >=20 > > Ok, looking at that part it appears it would do exactly what you need=
.
>=20 > > >=20 > > But I've noticed they have only 1 in stock? Sometimes when you see lo=
w
>=20 > > stock counts like that, it's a sign of stay away... ;) >=20 > Ah, didn't notice that, thanks... >=20 >=20 >=20 > Here's a "DUAL P-CHANNEL ENHANCEMENT MODE FIELD EFFECT TRANSISTOR" in sto=
ck.
>=20 >=20 >=20 > <http://www.digikey.com/product-detail/en/DMP2240UDM-7/DMP2240UDMDICT-ND/=
1774243>
>=20 >=20 >=20 > About .1=CE=A9 Rds(on) @ 4.5v, which is good for me. >=20 >=20 >=20 > The pinouts also look like they'll be convenient for my application.=20 >=20 > More than the single-fet packages.
just remember that now that you have turned things up side down the Vgs is = you supply - Vol of your controller So if you supply your LEDs with 1.8V it'll be at most 1.8V =20 -Lasse -Lasse
Reply by Daniel Pitts March 11, 20142014-03-11
On 3/11/14 4:39 PM, Maynard A. Philbrook Jr. wrote:
> In article <71JTu.1700$KZ7.1308@fx32.iad>, > newsgroup.nospam@virtualinfinity.net says... >> Looking more carefully, I think I might need P-Channel MOSFETS, and then >> I'd need to invert my logic driver (replace the 74HC238 with a 74HC138). >> If I used an N-Channel, I don't think I could get the Vgs to be high >> enough. >> >> I was thinking this part: >> http://www.digikey.com/product-detail/en/SCH1302-TL-E/869-1194-1-ND/2165672 >> >> Then I would have the source connected to +5v, the drain connected to to >> my load, and the gate connected to the output of the 1-of-8. >> >> The load is some parallel LEDs (each in series with a >> constant-current-sink), so the drain voltage could vary between +1.8v up >> to +5.0v, The current can vary from 10mA to 1200mA. Depending on how >> many parallel LEDs are enabled. >> >> Does this sound like it would work? >> >> Thanks, >> Daniel. >> > Ok, looking at that part it appears it would do exactly what you need. > > But I've noticed they have only 1 in stock? Sometimes when you see low > stock counts like that, it's a sign of stay away... ;)
Ah, didn't notice that, thanks... Here's a "DUAL P-CHANNEL ENHANCEMENT MODE FIELD EFFECT TRANSISTOR" in stock. <http://www.digikey.com/product-detail/en/DMP2240UDM-7/DMP2240UDMDICT-ND/1774243> About .1&Omega; Rds(on) @ 4.5v, which is good for me. The pinouts also look like they'll be convenient for my application. More than the single-fet packages.
Reply by Daniel Pitts March 11, 20142014-03-11
On 3/11/14 2:35 PM, Glenn wrote:
> On 11/03/14 19.58, Daniel Pitts wrote: > ... >> The load is some parallel LEDs (each in series with a >> constant-current-sink), so the drain voltage could vary between +1.8v up >> to +5.0v, The current can vary from 10mA to 1200mA. Depending on how >> many parallel LEDs are enabled. >> >> Does this sound like it would work? >> >> Thanks, >> Daniel. > > You could control the constant-current-sink directly? > > E.g. by sorting the gate/source or base/emitter by a small MOSFET (2N7002)?
I'm multiplexing 24*8 LEDs. The "current sink" is 3 8-bit shift-register constant-current-sink (TLC5196) which power between 0 to 24 LEDs in parallel (depending). The MOSFET is a "row enable" to choose between 1 of 8 sets of 24 LEDS. So in total, I'm powering 196 LEDs, just 24 at a time.
> > Double 2N7002: > > 2N7002DWA-7DICT-ND > (digikey) > > /Glenn
Reply by Maynard A. Philbrook Jr. March 11, 20142014-03-11
In article <71JTu.1700$KZ7.1308@fx32.iad>, 
newsgroup.nospam@virtualinfinity.net says...
> Looking more carefully, I think I might need P-Channel MOSFETS, and then > I'd need to invert my logic driver (replace the 74HC238 with a 74HC138). > If I used an N-Channel, I don't think I could get the Vgs to be high > enough. > > I was thinking this part: > http://www.digikey.com/product-detail/en/SCH1302-TL-E/869-1194-1-ND/2165672 > > Then I would have the source connected to +5v, the drain connected to to > my load, and the gate connected to the output of the 1-of-8. > > The load is some parallel LEDs (each in series with a > constant-current-sink), so the drain voltage could vary between +1.8v up > to +5.0v, The current can vary from 10mA to 1200mA. Depending on how > many parallel LEDs are enabled. > > Does this sound like it would work? > > Thanks, > Daniel. >
Ok, looking at that part it appears it would do exactly what you need. But I've noticed they have only 1 in stock? Sometimes when you see low stock counts like that, it's a sign of stay away... ;) Jamie
Reply by Glenn March 11, 20142014-03-11
On 11/03/14 19.58, Daniel Pitts wrote:
...
> The load is some parallel LEDs (each in series with a > constant-current-sink), so the drain voltage could vary between +1.8v up > to +5.0v, The current can vary from 10mA to 1200mA. Depending on how > many parallel LEDs are enabled. > > Does this sound like it would work? > > Thanks, > Daniel.
You could control the constant-current-sink directly? E.g. by sorting the gate/source or base/emitter by a small MOSFET (2N7002)? Double 2N7002: 2N7002DWA-7DICT-ND (digikey) /Glenn
Reply by Daniel Pitts March 11, 20142014-03-11
On 3/10/14 11:39 AM, Daniel Pitts wrote:
> On 3/9/14 1:43 PM, Phil Hobbs wrote: >> On 3/9/2014 4:33 PM, Daniel Pitts wrote: >>> On 3/9/14 1:25 PM, Daniel Pitts wrote: >>>> On 3/9/14 12:39 PM, Phil Hobbs wrote: >>>>> On 3/9/2014 3:21 PM, Daniel Pitts wrote: >>>>>> I'm going to be using ULN2803APG in a design of mine, and I just >>>>>> noticed >>>>>> on the datasheet [1] that Vce(sat) typ is 1.3 at 350ma. >>>>>> >>>>>> Does that mean that the voltage drop across the Vce will typically be >>>>>> 1.3 at that current? >>>>> >>>>> 'Fraid so. It's a Darlington, intended for high voltage outputs. >>>> >>>> That was what I was afraid of. I should still have enough headroom, but >>>> that is wasting a bit of power... >>>> >>>> Any suggestion on a replacement part? I need to enable/disable 8 lines >>>> with logic level inputs. >>>> >>>> It can even be a 1-of-8 decoder I just need it to support up to 1A, and >>>> I'd like it to be relatively cheap. My load is going to 24 parallel >>>> LEDs (3.3v Vf for 16 of them, 1.8 Vf for 8), each with a constant >>>> current sink in line with them. >>>> >>>> Cheap, for me, is < $1.25 for 8-in-8-out, or < $2.0 for 3-in-8-out >>>> (since I'd be able to replace two chips in that case). >>> I should add, I'm talking about price for quantities < 20 at the moment, >>> not including shipping. >>> >> >> You'd think that somebody would make a MOSFET version, but AFAIK they >> don't. > Maybe what I really need is just 8 Enhancment-Mode N-Channel MOSFETs? > > Such as 8 of these? > http://www.digikey.com/product-detail/en/IRLML2060TRPBF/IRLML2060TRPBFCT-ND/2271917
Looking more carefully, I think I might need P-Channel MOSFETS, and then I'd need to invert my logic driver (replace the 74HC238 with a 74HC138). If I used an N-Channel, I don't think I could get the Vgs to be high enough. I was thinking this part: http://www.digikey.com/product-detail/en/SCH1302-TL-E/869-1194-1-ND/2165672 Then I would have the source connected to +5v, the drain connected to to my load, and the gate connected to the output of the 1-of-8. The load is some parallel LEDs (each in series with a constant-current-sink), so the drain voltage could vary between +1.8v up to +5.0v, The current can vary from 10mA to 1200mA. Depending on how many parallel LEDs are enabled. Does this sound like it would work? Thanks, Daniel.
Reply by miso March 11, 20142014-03-11
Glenn wrote:

>> That's fine for my case. I'm using an 74HCT238 @ 5V to run the thing. >
I was thinking more about the strength of the fet if it is on the high side of the threshold spec. Just something to consider.
Reply by Jim Thompson March 10, 20142014-03-10
On Mon, 10 Mar 2014 11:48:26 -0800, Robert Baer
<robertbaer@localnet.com> wrote:

>Phil Hobbs wrote: >> On 3/9/2014 3:21 PM, Daniel Pitts wrote: >>> I'm going to be using ULN2803APG in a design of mine, and I just noticed >>> on the datasheet [1] that Vce(sat) typ is 1.3 at 350ma. >>> >>> Does that mean that the voltage drop across the Vce will typically be >>> 1.3 at that current? >> >> 'Fraid so. It's a Darlington, intended for high voltage outputs. >> >> Cheers >> >> Phil Hobbs >> > Check; it is impossible to saturate a darlington.
The main device... but the driver device saturates. ...Jim Thompson -- | James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
Reply by Robert Baer March 10, 20142014-03-10
Phil Hobbs wrote:
> On 3/9/2014 3:21 PM, Daniel Pitts wrote: >> I'm going to be using ULN2803APG in a design of mine, and I just noticed >> on the datasheet [1] that Vce(sat) typ is 1.3 at 350ma. >> >> Does that mean that the voltage drop across the Vce will typically be >> 1.3 at that current? > > 'Fraid so. It's a Darlington, intended for high voltage outputs. > > Cheers > > Phil Hobbs >
Check; it is impossible to saturate a darlington.