On Friday, September 30, 2022 at 9:06:34 PM UTC-4, bitrex wrote:
> On 9/30/2022 7:33 PM, Ricky wrote: 
> > On Friday, September 30, 2022 at 9:34:16 AM UTC-4, bitrex wrote: 
> >> On 9/29/2022 1:09 AM, Ricky wrote: 
> >>> On Thursday, September 29, 2022 at 12:44:48 AM UTC-4, bitrex wrote: 
> >>>> On 9/26/2022 4:09 PM, Ricky wrote: 
> >>>>> A customer wants me to redesign a board to eliminate the production bottlenecks. They also want all IP so they can make the boards themselves if my company is unable to. I'm fine with that, but I'd like to have some means of assurance they won't make boards without my royalty being respected. 
> >>>>> 
> >>>>> The board has an FPGA which contains the "magic", an analog path, and a digital path to the outside world. The digital path needs a 3.3V/5V interface. There are two opamps that serve as filters with gain. There is a need for several (3-4) LDOs. 
> >>>>> 
> >>>>> I've found a couple of chips from Greenpaks that could help here. One is a "Programmable Mixed-Signal Matrix" which could replace the opamps and provide a configurable gain using the programmable "rheostat". Another has four LDOs which would be useful and *might* be able to serve as the level shifter. 
> >>>>> 
> >>>>> I'm waiting to hear back from someone from Renesas, who can discuss this with me, or a disti FAE. There are a lot of questions about how to turn these into a custom part number to meet my needs. 
> >>>>> 
> >>>>> Anyone have experience with using these in production? 
> >>>>> 
> >>>> Please let me know if you're able to get anyone and how it goes, It 
> >>>> seems like the Renesas acquisition of Dialog's lineup has been sloppy as 
> >>>> hell 
> >>> 
> >>> I thought someone here has used the Greenpak parts before. Maybe it was another group. 
> >>> 
> >> Nah it was me (though there another guy here who's used them also, "BoB" 
> >> I believe.) 
> >> 
> >> I've done a number of designs with them in the past couple years so if 
> >> you have any questions I might be able to help. 
> > 
> > I'm a bit unclear about the programming. When they have an I2C port, does that mean they can be programmed via that? I'm looking at the SLG47004 and SLG46580. Can the SLG46580 LDOs be set to power up to a given voltage? I don't see a means of sequencing the power up of the LDO, other than perhaps through the I2C port. They talk about writing to registers that reflect the non-volatile memory. It's not clear if there are registers that are not loaded from the NVM.
> You can set the LDOs to power up to any of the 32 steps between 0.9 and 
> 4.35. They also have a "pin" for enable/disable, and there's a "pin" to 
> select between two different voltages you pick, or to pass the input 
> directly to the output, depending on how that "pin" is configured. 
> 
> The SLG46580 also has an 8 state ASM, you could probably use that for 
> power sequencing if it's not otherwise needed..
> > Do all of their parts have factory programmed NVM? Do their parts let you configure everything from the I2C port, or do the parts need to be replaced when reprogrammed?
> There's two layers to a program, an NVM layer and a RAM register layer. 
> What's programmed on the factory on the OTP units (a couple are NVM-MTP 
> but I've not used them) is the NVM layer, but at power-up that data is 
> pulled into the registers in RAM. 
> 
> After that it's what's in the RAM that determines how the chip operates, 
> so you can write those via i2c once it's running and re-configure it as 
> you like. 
> 
> See page 57 in the PDF: 
> 
> <https://www.renesas.com/us/en/document/mat/greenpak-cookbook> 
> 
> There's a tool in the designer that lets you take a snapshot of two 
> designs for a given device and it will spit out a set of difference 
> commands you can send over i2c that will take it to the next 
> configuration from the previous.
> > Whew! I started typing this around noon and have been slammed all afternoon! 
> > 
> > Looks like the respin is going to be rewarding. They are talking about 18,000 units next year and more in the following two years! I just need to design it with an FPGA that I can actually buy, not to mention the other parts on the board. I hope the availability of the Greenpak units is not too bad. 

I've been looking for an image of a pre-programmed part, or any Greenpak part, for that matter.  Can't find one.  The data sheet says the markings are pretty basic, with a "part ID", a date/lot code, assembly site and revision code.  I think the part ID and revision code refer to the customer's design.  It doesn't look like there's room for a logo or other information identifying the manufacturer.  

Do you have a photo of a pre-programmed part, you can share? 

-- 

Rick C.

++- Get 1,000 miles of free Supercharging
++- Tesla referral code - https://ts.la/richard11209

On Friday, September 30, 2022 at 9:06:34 PM UTC-4, bitrex wrote:
> On 9/30/2022 7:33 PM, Ricky wrote: 
> > On Friday, September 30, 2022 at 9:34:16 AM UTC-4, bitrex wrote: 
> >> On 9/29/2022 1:09 AM, Ricky wrote: 
> >>> On Thursday, September 29, 2022 at 12:44:48 AM UTC-4, bitrex wrote: 
> >>>> On 9/26/2022 4:09 PM, Ricky wrote: 
> >>>>> A customer wants me to redesign a board to eliminate the production bottlenecks. They also want all IP so they can make the boards themselves if my company is unable to. I'm fine with that, but I'd like to have some means of assurance they won't make boards without my royalty being respected. 
> >>>>> 
> >>>>> The board has an FPGA which contains the "magic", an analog path, and a digital path to the outside world. The digital path needs a 3.3V/5V interface. There are two opamps that serve as filters with gain. There is a need for several (3-4) LDOs. 
> >>>>> 
> >>>>> I've found a couple of chips from Greenpaks that could help here. One is a "Programmable Mixed-Signal Matrix" which could replace the opamps and provide a configurable gain using the programmable "rheostat". Another has four LDOs which would be useful and *might* be able to serve as the level shifter. 
> >>>>> 
> >>>>> I'm waiting to hear back from someone from Renesas, who can discuss this with me, or a disti FAE. There are a lot of questions about how to turn these into a custom part number to meet my needs. 
> >>>>> 
> >>>>> Anyone have experience with using these in production? 
> >>>>> 
> >>>> Please let me know if you're able to get anyone and how it goes, It 
> >>>> seems like the Renesas acquisition of Dialog's lineup has been sloppy as 
> >>>> hell 
> >>> 
> >>> I thought someone here has used the Greenpak parts before. Maybe it was another group. 
> >>> 
> >> Nah it was me (though there another guy here who's used them also, "BoB" 
> >> I believe.) 
> >> 
> >> I've done a number of designs with them in the past couple years so if 
> >> you have any questions I might be able to help. 
> > 
> > I'm a bit unclear about the programming. When they have an I2C port, does that mean they can be programmed via that? I'm looking at the SLG47004 and SLG46580. Can the SLG46580 LDOs be set to power up to a given voltage? I don't see a means of sequencing the power up of the LDO, other than perhaps through the I2C port. They talk about writing to registers that reflect the non-volatile memory. It's not clear if there are registers that are not loaded from the NVM.
> You can set the LDOs to power up to any of the 32 steps between 0.9 and 
> 4.35. They also have a "pin" for enable/disable, and there's a "pin" to 
> select between two different voltages you pick, or to pass the input 
> directly to the output, depending on how that "pin" is configured. 
> 
> The SLG46580 also has an 8 state ASM, you could probably use that for 
> power sequencing if it's not otherwise needed..
> > Do all of their parts have factory programmed NVM? Do their parts let you configure everything from the I2C port, or do the parts need to be replaced when reprogrammed?
> There's two layers to a program, an NVM layer and a RAM register layer. 
> What's programmed on the factory on the OTP units (a couple are NVM-MTP 
> but I've not used them) is the NVM layer, but at power-up that data is 
> pulled into the registers in RAM. 
> 
> After that it's what's in the RAM that determines how the chip operates, 
> so you can write those via i2c once it's running and re-configure it as 
> you like. 
> 
> See page 57 in the PDF: 
> 
> <https://www.renesas.com/us/en/document/mat/greenpak-cookbook> 
> 
> There's a tool in the designer that lets you take a snapshot of two 
> designs for a given device and it will spit out a set of difference 
> commands you can send over i2c that will take it to the next 
> configuration from the previous.
> > Whew! I started typing this around noon and have been slammed all afternoon! 
> > 
> > Looks like the respin is going to be rewarding. They are talking about 18,000 units next year and more in the following two years! I just need to design it with an FPGA that I can actually buy, not to mention the other parts on the board. I hope the availability of the Greenpak units is not too bad. 
> >

I've tried emailing several contacts listed at the Renesas site, a couple at Avnet and a couple at Renesas.  No responses.  Have you spoken to anyone about the GreenPak devices?  I'd like to spend a few minutes talking to an FAE.  

-- 

Rick C.

+-+ Get 1,000 miles of free Supercharging
+-+ Tesla referral code - https://ts.la/richard11209

On Friday, September 30, 2022 at 9:06:34 PM UTC-4, bitrex wrote:
> On 9/30/2022 7:33 PM, Ricky wrote: 
> > On Friday, September 30, 2022 at 9:34:16 AM UTC-4, bitrex wrote: 
> >> On 9/29/2022 1:09 AM, Ricky wrote: 
> >>> On Thursday, September 29, 2022 at 12:44:48 AM UTC-4, bitrex wrote: 
> >>>> On 9/26/2022 4:09 PM, Ricky wrote: 
> >>>>> A customer wants me to redesign a board to eliminate the production bottlenecks. They also want all IP so they can make the boards themselves if my company is unable to. I'm fine with that, but I'd like to have some means of assurance they won't make boards without my royalty being respected. 
> >>>>> 
> >>>>> The board has an FPGA which contains the "magic", an analog path, and a digital path to the outside world. The digital path needs a 3.3V/5V interface. There are two opamps that serve as filters with gain. There is a need for several (3-4) LDOs. 
> >>>>> 
> >>>>> I've found a couple of chips from Greenpaks that could help here. One is a "Programmable Mixed-Signal Matrix" which could replace the opamps and provide a configurable gain using the programmable "rheostat". Another has four LDOs which would be useful and *might* be able to serve as the level shifter. 
> >>>>> 
> >>>>> I'm waiting to hear back from someone from Renesas, who can discuss this with me, or a disti FAE. There are a lot of questions about how to turn these into a custom part number to meet my needs. 
> >>>>> 
> >>>>> Anyone have experience with using these in production? 
> >>>>> 
> >>>> Please let me know if you're able to get anyone and how it goes, It 
> >>>> seems like the Renesas acquisition of Dialog's lineup has been sloppy as 
> >>>> hell 
> >>> 
> >>> I thought someone here has used the Greenpak parts before. Maybe it was another group. 
> >>> 
> >> Nah it was me (though there another guy here who's used them also, "BoB" 
> >> I believe.) 
> >> 
> >> I've done a number of designs with them in the past couple years so if 
> >> you have any questions I might be able to help. 
> > 
> > I'm a bit unclear about the programming. When they have an I2C port, does that mean they can be programmed via that? I'm looking at the SLG47004 and SLG46580. Can the SLG46580 LDOs be set to power up to a given voltage? I don't see a means of sequencing the power up of the LDO, other than perhaps through the I2C port. They talk about writing to registers that reflect the non-volatile memory. It's not clear if there are registers that are not loaded from the NVM.
> You can set the LDOs to power up to any of the 32 steps between 0.9 and 
> 4.35. They also have a "pin" for enable/disable, and there's a "pin" to 
> select between two different voltages you pick, or to pass the input 
> directly to the output, depending on how that "pin" is configured. 
> 
> The SLG46580 also has an 8 state ASM, you could probably use that for 
> power sequencing if it's not otherwise needed..

Yeah, both of the parts I've found so far have lots of stuff I won't use.  The logic doesn't do much for me and the comparators are not of much value.  If only they combined the op amps with the LDOs and threw in a 5V to 3V level shifter.  :) 


> > Do all of their parts have factory programmed NVM? Do their parts let you configure everything from the I2C port, or do the parts need to be replaced when reprogrammed?
> There's two layers to a program, an NVM layer and a RAM register layer. 
> What's programmed on the factory on the OTP units (a couple are NVM-MTP 
> but I've not used them) is the NVM layer, but at power-up that data is 
> pulled into the registers in RAM. 
> 
> After that it's what's in the RAM that determines how the chip operates, 
> so you can write those via i2c once it's running and re-configure it as 
> you like. 

I kinda figured that.  Is there a way to lock bits of the configuration so they can't be read or changed?  I was hoping when the factory programmed it there would be some level of security.  I suppose it can be a good thing to support field upgrades though. 


> See page 57 in the PDF: 
> 
> <https://www.renesas.com/us/en/document/mat/greenpak-cookbook> 
> 
> There's a tool in the designer that lets you take a snapshot of two 
> designs for a given device and it will spit out a set of difference 
> commands you can send over i2c that will take it to the next 
> configuration from the previous.

That's interesting.  It's not so much that I will be loading multiple designs, other than adjusting the rheostat in the opamp part.  

-- 

Rick C.

+-- Get 1,000 miles of free Supercharging
+-- Tesla referral code - https://ts.la/richard11209

On 9/30/2022 7:33 PM, Ricky wrote:
> On Friday, September 30, 2022 at 9:34:16 AM UTC-4, bitrex wrote:
>> On 9/29/2022 1:09 AM, Ricky wrote:
>>> On Thursday, September 29, 2022 at 12:44:48 AM UTC-4, bitrex wrote:
>>>> On 9/26/2022 4:09 PM, Ricky wrote:
>>>>> A customer wants me to redesign a board to eliminate the production bottlenecks. They also want all IP so they can make the boards themselves if my company is unable to. I'm fine with that, but I'd like to have some means of assurance they won't make boards without my royalty being respected.
>>>>>
>>>>> The board has an FPGA which contains the "magic", an analog path, and a digital path to the outside world. The digital path needs a 3.3V/5V interface. There are two opamps that serve as filters with gain. There is a need for several (3-4) LDOs.
>>>>>
>>>>> I've found a couple of chips from Greenpaks that could help here. One is a "Programmable Mixed-Signal Matrix" which could replace the opamps and provide a configurable gain using the programmable "rheostat". Another has four LDOs which would be useful and *might* be able to serve as the level shifter.
>>>>>
>>>>> I'm waiting to hear back from someone from Renesas, who can discuss this with me, or a disti FAE. There are a lot of questions about how to turn these into a custom part number to meet my needs.
>>>>>
>>>>> Anyone have experience with using these in production?
>>>>>
>>>> Please let me know if you're able to get anyone and how it goes, It
>>>> seems like the Renesas acquisition of Dialog's lineup has been sloppy as
>>>> hell
>>>
>>> I thought someone here has used the Greenpak parts before. Maybe it was another group.
>>>
>> Nah it was me (though there another guy here who's used them also, "BoB"
>> I believe.)
>>
>> I've done a number of designs with them in the past couple years so if
>> you have any questions I might be able to help.
> 
> I'm a bit unclear about the programming.  When they have an I2C port, does that mean they can be programmed via that?  I'm looking at the SLG47004 and SLG46580.  Can the SLG46580 LDOs be set to power up to a given voltage?  I don't see a means of sequencing the power up of the LDO, other than perhaps through the I2C port.  They talk about writing to registers that reflect the non-volatile memory.  It's not clear if there are registers that are not loaded from the NVM.

You can set the LDOs to power up to any of the 32 steps between 0.9 and 
4.35. They also have a "pin" for enable/disable, and there's a "pin" to 
select between two different voltages you pick, or to pass the input 
directly to the output, depending on how that "pin" is configured.

The SLG46580 also has an 8 state ASM, you could probably use that for 
power sequencing if it's not otherwise needed..

> Do all of their parts have factory programmed NVM?  Do their parts let you configure everything from the I2C port, or do the parts need to be replaced when reprogrammed?

There's two layers to a program, an NVM layer and a RAM register layer. 
What's programmed on the factory on the OTP units (a couple are NVM-MTP 
but I've not used them) is the NVM layer, but at power-up that data is 
pulled into the registers in RAM.

After that it's what's in the RAM that determines how the chip operates, 
so you can write those via i2c once it's running and re-configure it as 
you like.

See page 57 in the PDF:

<https://www.renesas.com/us/en/document/mat/greenpak-cookbook>

There's a tool in the designer that lets you take a snapshot of two 
designs for a given device and it will spit out a set of difference 
commands you can send over i2c that will take it to the next 
configuration from the previous.

> Whew!  I started typing this around noon and have been slammed all afternoon!
> 
> Looks like the respin is going to be rewarding.  They are talking about 18,000 units next year and more in the following two years!  I just need to design it with an FPGA that I can actually buy, not to mention the other parts on the board.  I hope the availability of the Greenpak units is not too bad.
>

On Friday, September 30, 2022 at 9:52:15 AM UTC-4, bitrex wrote:
> On 9/29/2022 1:09 AM, Ricky wrote:
> > On Thursday, September 29, 2022 at 12:44:48 AM UTC-4, bitrex wrote: 
> >> On 9/26/2022 4:09 PM, Ricky wrote: 
> >>> A customer wants me to redesign a board to eliminate the production bottlenecks. They also want all IP so they can make the boards themselves if my company is unable to. I'm fine with that, but I'd like to have some means of assurance they won't make boards without my royalty being respected. 
> >>> 
> >>> The board has an FPGA which contains the "magic", an analog path, and a digital path to the outside world. The digital path needs a 3.3V/5V interface. There are two opamps that serve as filters with gain. There is a need for several (3-4) LDOs. 
> >>> 
> >>> I've found a couple of chips from Greenpaks that could help here. One is a "Programmable Mixed-Signal Matrix" which could replace the opamps and provide a configurable gain using the programmable "rheostat". Another has four LDOs which would be useful and *might* be able to serve as the level shifter. 
> >>> 
> >>> I'm waiting to hear back from someone from Renesas, who can discuss this with me, or a disti FAE. There are a lot of questions about how to turn these into a custom part number to meet my needs. 
> >>> 
> >>> Anyone have experience with using these in production? 
> >>> 
> >> Please let me know if you're able to get anyone and how it goes, It 
> >> seems like the Renesas acquisition of Dialog's lineup has been sloppy as 
> >> hell 
> > 
> > I thought someone here has used the Greenpak parts before. Maybe it was another group. 
> >
> BTW I noticed you talked about level shifting, a number of GreenPak 
> parts have dual supply pins and can intrinsically level shift between 
> 3.3 and 5V, you put your 3.3 I/O on one set of color-coded pins and the 
> 5V on the other set, easily done.

I don't think the parts that are otherwise useful have the dual voltages.  I just need 3.3V I/O that are 5V tolerant.  I'm not finding that on the parts that otherwise look interesting. 

-- 

Rick C.

-++ Get 1,000 miles of free Supercharging
-++ Tesla referral code - https://ts.la/richard11209

On Friday, September 30, 2022 at 9:34:16 AM UTC-4, bitrex wrote:
> On 9/29/2022 1:09 AM, Ricky wrote: 
> > On Thursday, September 29, 2022 at 12:44:48 AM UTC-4, bitrex wrote: 
> >> On 9/26/2022 4:09 PM, Ricky wrote: 
> >>> A customer wants me to redesign a board to eliminate the production bottlenecks. They also want all IP so they can make the boards themselves if my company is unable to. I'm fine with that, but I'd like to have some means of assurance they won't make boards without my royalty being respected. 
> >>> 
> >>> The board has an FPGA which contains the "magic", an analog path, and a digital path to the outside world. The digital path needs a 3.3V/5V interface. There are two opamps that serve as filters with gain. There is a need for several (3-4) LDOs. 
> >>> 
> >>> I've found a couple of chips from Greenpaks that could help here. One is a "Programmable Mixed-Signal Matrix" which could replace the opamps and provide a configurable gain using the programmable "rheostat". Another has four LDOs which would be useful and *might* be able to serve as the level shifter. 
> >>> 
> >>> I'm waiting to hear back from someone from Renesas, who can discuss this with me, or a disti FAE. There are a lot of questions about how to turn these into a custom part number to meet my needs. 
> >>> 
> >>> Anyone have experience with using these in production? 
> >>> 
> >> Please let me know if you're able to get anyone and how it goes, It 
> >> seems like the Renesas acquisition of Dialog's lineup has been sloppy as 
> >> hell 
> > 
> > I thought someone here has used the Greenpak parts before. Maybe it was another group. 
> >
> Nah it was me (though there another guy here who's used them also, "BoB" 
> I believe.) 
> 
> I've done a number of designs with them in the past couple years so if 
> you have any questions I might be able to help.

I'm a bit unclear about the programming.  When they have an I2C port, does that mean they can be programmed via that?  I'm looking at the SLG47004 and SLG46580.  Can the SLG46580 LDOs be set to power up to a given voltage?  I don't see a means of sequencing the power up of the LDO, other than perhaps through the I2C port.  They talk about writing to registers that reflect the non-volatile memory.  It's not clear if there are registers that are not loaded from the NVM.  

Do all of their parts have factory programmed NVM?  Do their parts let you configure everything from the I2C port, or do the parts need to be replaced when reprogrammed? 

Whew!  I started typing this around noon and have been slammed all afternoon!  

Looks like the respin is going to be rewarding.  They are talking about 18,000 units next year and more in the following two years!  I just need to design it with an FPGA that I can actually buy, not to mention the other parts on the board.  I hope the availability of the Greenpak units is not too bad. 

-- 

Rick C.

-+- Get 1,000 miles of free Supercharging
-+- Tesla referral code - https://ts.la/richard11209

On 9/29/2022 1:09 AM, Ricky wrote:
> On Thursday, September 29, 2022 at 12:44:48 AM UTC-4, bitrex wrote:
>> On 9/26/2022 4:09 PM, Ricky wrote:
>>> A customer wants me to redesign a board to eliminate the production bottlenecks. They also want all IP so they can make the boards themselves if my company is unable to. I'm fine with that, but I'd like to have some means of assurance they won't make boards without my royalty being respected.
>>>
>>> The board has an FPGA which contains the "magic", an analog path, and a digital path to the outside world. The digital path needs a 3.3V/5V interface. There are two opamps that serve as filters with gain. There is a need for several (3-4) LDOs.
>>>
>>> I've found a couple of chips from Greenpaks that could help here. One is a "Programmable Mixed-Signal Matrix" which could replace the opamps and provide a configurable gain using the programmable "rheostat". Another has four LDOs which would be useful and *might* be able to serve as the level shifter.
>>>
>>> I'm waiting to hear back from someone from Renesas, who can discuss this with me, or a disti FAE. There are a lot of questions about how to turn these into a custom part number to meet my needs.
>>>
>>> Anyone have experience with using these in production?
>>>
>> Please let me know if you're able to get anyone and how it goes, It
>> seems like the Renesas acquisition of Dialog's lineup has been sloppy as
>> hell
> 
> I thought someone here has used the Greenpak parts before.  Maybe it was another group.
> 

BTW I noticed you talked about level shifting, a number of GreenPak 
parts have dual supply pins and can intrinsically level shift between 
3.3 and 5V, you put your 3.3 I/O on one set of color-coded pins and the 
5V on the other set, easily done.

On 9/29/2022 1:09 AM, Ricky wrote:
> On Thursday, September 29, 2022 at 12:44:48 AM UTC-4, bitrex wrote:
>> On 9/26/2022 4:09 PM, Ricky wrote:
>>> A customer wants me to redesign a board to eliminate the production bottlenecks. They also want all IP so they can make the boards themselves if my company is unable to. I'm fine with that, but I'd like to have some means of assurance they won't make boards without my royalty being respected.
>>>
>>> The board has an FPGA which contains the "magic", an analog path, and a digital path to the outside world. The digital path needs a 3.3V/5V interface. There are two opamps that serve as filters with gain. There is a need for several (3-4) LDOs.
>>>
>>> I've found a couple of chips from Greenpaks that could help here. One is a "Programmable Mixed-Signal Matrix" which could replace the opamps and provide a configurable gain using the programmable "rheostat". Another has four LDOs which would be useful and *might* be able to serve as the level shifter.
>>>
>>> I'm waiting to hear back from someone from Renesas, who can discuss this with me, or a disti FAE. There are a lot of questions about how to turn these into a custom part number to meet my needs.
>>>
>>> Anyone have experience with using these in production?
>>>
>> Please let me know if you're able to get anyone and how it goes, It
>> seems like the Renesas acquisition of Dialog's lineup has been sloppy as
>> hell
> 
> I thought someone here has used the Greenpak parts before.  Maybe it was another group.
> 

Nah it was me (though there another guy here who's used them also, "BoB" 
I believe.)

I've done a number of designs with them in the past couple years so if 
you have any questions I might be able to help.