Predictika · Programmable DC power supplies for test benches

An AI sales agent's DC power supply order, checked before it goes out

An AI sales agent prepares an order for Sorensen SG Series programmable DC power supplies for a test bench. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the SG Series datasheet and returns a verdict, every reason and the figures it computed.

Scope: AMETEK Programmable Power Sorensen SG Series (SGA analog and SGI intelligent) DC power supplies, 10 to 600 V and 4 to 30 kW per chassis, paralleled up to 150 kW. See everything this validator covers

This validator models orders for the Sorensen SG Series as listed in the public SG Series datasheet, not just the scenarios shown below.

  • Models — the SGA (analog control) and SGI (intelligent control) in the 15 voltage ratings from 10 V to 600 V, with one to six 5 kW power modules (4 kW modules on the 10 V and 15 V ratings) in a 3U or 6U chassis, and the output current the datasheet lists for each combination.
  • Chassis rules — the 6U chassis is offered as a single unit only from 60 V to 600 V; at 10 to 50 V the larger power steps come only by paralleling 3U supplies.
  • Options and accessories — input voltage options C, D and E, remote control options (IEEE-488.2, Ethernet, isolated analog control, shaft locks), the dust filter, 400 Hz input and power factor modifications, paralleling cables and rack slides.
  • System rules — up to 5 units in parallel and 2 in series, output float voltage, 0 to 50 degrees C operation, and the altitude derating above 5,000 feet; the test voltage, current and power needed are compared with what the order gives.
  • Figures checked — the total power and total current the order states are compared with units times the datasheet figures.

Any order for these products that touches these rules is in scope — not only the scenarios pictured here. Not covered: prices (none are modeled, so there is no price build-up or line-item quote), the 800 V models, other AMETEK Programmable Power product lines, and any change AMETEK has made to the SG Series since the 2011 datasheet.

The AI sales agentBuilds the power supply order from the conversation and says it is ready.
Predictika validatorChecks every rule at once and returns the verdict, each finding, and the computed figures.

7 scenarios in 3 groups · each group runs from the agent’s first attempt to the outcome · click one to switch

What the agent proposes

The AI sales agent says

Illustrative agent statement. The validator receives the structured request below.

Proposing agent: AI sales agent (scripted)

Before sending, the agent must call Recorded run

Predictika validator


      

        
        
        
      

      
Waiting for the call.

      
    

  

  
  

    
Where rules-based CPQ gets it wrong
Click to expand
    
Our assessment of what a typical rules-based CPQ (product rules, price rules and validation rules, as in Salesforce CPQ) does with the same cases. It is a judgment about effort and reliability, not a claim that nothing can be built.

    
What the validator caughtTypical rules-based CPQPredictika
A 6U chassis ordered at a voltage where only paralleled 3U supplies reach that powerOnly with care
The single-chassis table has gaps at 10 to 50 V, so every voltage and power step needs its own availability row.		Voltage and power step are compared with the datasheet table on every call.
Options that depend on the chassis, the control type or the input voltage (dust filter, 400 Hz, shaft locks)Only with care
Each option has its own exclusion by chassis or control, so the rules multiply and are easy to miss.		Every option is compared with the chassis, control and input choice on every call.
Total power and current stated by the agent that do not equal units times the datasheet figuresOnly with workarounds
The stated figure is usually free text or a separate field, so it takes an extra rule that multiplies and compares.		The stated figures are part of the request and are compared with the computed figures on every call.
A test load the order cannot deliver at the site altitudeOnly with workarounds
Needs a derating calculation outside the product rules.		Voltage, current and power need are checked together with the altitude derating on every call.
Every problem returned in one answerOnly with workarounds
Rules tend to fire as a step is completed, so a rep often fixes one error and meets the next.		All findings come back together, whatever order the choices arrived in.

  

  

    
What this validator catches: findings from its test cases
Click to expand
    
Every rule is checked on every call, in any order the choices arrive; one problem never hides another. Blocking findings stop the order; approval findings need a person to sign off; everything else is computed.

    
1The datasheet gets this voltage and power in a 6U chassis only by paralleling 3U supplies, so it is not offered as a single unit.
2120 V is not a voltage rating this validator covers; SG ratings modeled are 10, 15, 20, 30, 40, 50, 60, 80, 100, 160, 200, 250, 330, 400 and 600 V.
3The power size step 7 is not one of the six steps (one to six power modules) of the SG series.
46 units in parallel; the datasheet allows up to 5 (contact the factory for more).
53 units in parallel need one paralleling cable per unit, but the order has 2 cable(s).
6The order has 1 paralleling cable(s) for a single unit.
73 units in series; the datasheet allows up to 2.
8The negative terminal would float 350 V from chassis potential; the datasheet allows within 300 V.
9The output floats 200 V from chassis potential and the datasheet recommends the isolated analog interface (option 1D); the customer must confirm.
10The shaft locks option (1E) is for the SGA series only.
11The factory-installed front panel dust filter (AJ) is for 3U units only.
12400 Hz input is standard on 3U units; the CV modification applies to 6U units only.
13The CV modification (400 Hz input) means the unit does not carry CE, CSA or UL marks; the customer must accept that.
14The CV modification is 400 Hz AC input at 208 VAC, which needs input option C.
15Passive power factor correction (PF) is for 40 V and 60 V to 600 V models; it is already included on 10 V to 30 V and 50 V models.
163U rack slides (K550,212-01) fit the 5, 10 and 15 kW models, not the 6U chassis.
176U rack slides (K550,213-01) fit the 20, 25 and 30 kW models, not the 3U chassis.
18The test needs 120 V but the supply is rated 100 V.
19The test needs 200 A but 1 unit(s) of this model give 150 A in total.
20The test needs 14,000 W but the supply gives 12,000 W after a 20 percent altitude derating.
21The datasheet lists operation at 0 to 50 degrees C; the site is at 55 degrees C.
22The order states 20 kW in total, but 1 unit(s) of 15 kW give 15 kW.
23The order states 160 A in total, but 1 unit(s) of 150 A give 150 A.
24The output floats 100 V from chassis potential and the datasheet recommends the isolated analog interface (option 1D); the customer must confirm.
250 units in parallel; the datasheet allows up to 5 (contact the factory for more).
26The test needs 100 A but 0 unit(s) of this model give 0 A in total.
27The test needs 8,000 W but the supply gives 0 W after a 0 percent altitude derating.

  

  

    
Scenarios and results
Click to expand
    
The recorded result of the validator for each scenario on this page: the request the AI agent sent, the verdict, and every finding, in the order they appear above.

    
1. Battery pack test bench: Two 330 V 30 kW SGI supplies in parallel with the 400 Hz modification

    
The agent's request. Please quote two Sorensen SGI supplies, 330 V and 30 kW each, with input option E (396 to 528 VAC) and Ethernet and RS-232C, with the 400 Hz input modification, with 2 paralleling cables. The test needs 300 V and 170 A at a site of 3,000 feet. Together they give 60 kW and 182 A. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • 400 Hz input needs the 208 V input option. The CV modification is 400 Hz AC input at 208 VAC, which needs input option C.
    • 
      
  • 400 Hz modification removes CE, CSA and UL marks. The CV modification (400 Hz input) means the unit does not carry CE, CSA or UL marks; the customer must accept that.
    • 
    

    
2. Battery pack test bench: The agent resubmits after fixing the errors

    
The agent's request. Please quote two Sorensen SGI supplies, 330 V and 30 kW each, with input option E (396 to 528 VAC) and Ethernet and RS-232C, with 2 paralleling cables. The test needs 300 V and 170 A at a site of 3,000 feet. Together they give 60 kW and 182 A. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Order summary. 2 unit(s) rated 330 V give 60 kW and 182 A in total.
    • 
    

    
3. Fuel cell stack test: One 10 V 16 kW SGI supply for a fuel cell stack test

    
The agent's request. Please quote one Sorensen SGI supply, 10 V and 16 kW each, with input option D (342 to 440 VAC) and no remote option. The test needs 8 V and 1,500 A at a site of 1,000 feet. Together they give 16 kW and 1,600 A. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Single 6U chassis not offered at this voltage. The datasheet gets this voltage and power in a 6U chassis only by paralleling 3U supplies, so it is not offered as a single unit.
    • 
    

    
4. Fuel cell stack test: The agent resubmits after fixing the errors

    
The agent's request. Please quote two Sorensen SGI supplies, 10 V and 8 kW each, with input option D (342 to 440 VAC) and no remote option, with 2 paralleling cables. The test needs 8 V and 1,500 A at a site of 1,000 feet. Together they give 16 kW and 1,600 A. It is ready to send.

    
Verdict: Allowed.

    
    
      
  • Order summary. 2 unit(s) rated 10 V give 16 kW and 1,600 A in total.
    • 
    

    
5. Floating test rig: One 200 V 15 kW SGA supply on a 400 V float

    
The agent's request. Please quote one Sorensen SGA supply, 200 V and 15 kW each, with input option D (342 to 440 VAC) and no remote option, with the output floating 400 V from chassis. The test needs 150 V and 60 A at a site of 1,000 feet. Together they give 15 kW and 75 A. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Output float voltage above 300 V. The negative terminal would float 400 V from chassis potential; the datasheet allows within 300 V.
    • 
      
  • Floating output without isolated analog control. The output floats 400 V from chassis potential and the datasheet recommends the isolated analog interface (option 1D); the customer must confirm.
    • 
    

    
6. Floating test rig: The agent resubmits after fixing the errors

    
The agent's request. Please quote one Sorensen SGA supply, 200 V and 15 kW each, with input option D (342 to 440 VAC) and no remote option, with the output floating 250 V from chassis. The test needs 150 V and 60 A at a site of 1,000 feet. Together they give 15 kW and 75 A. It is ready to send.

    
Verdict: Needs approval.

    
    
      
  • Floating output without isolated analog control. The output floats 250 V from chassis potential and the datasheet recommends the isolated analog interface (option 1D); the customer must confirm.
    • 
      
  • Order summary. 1 unit(s) rated 200 V give 15 kW and 75 A in total.
    • 
    

    
7. Mountain-site rack: One 100 V 15 kW SGI supply at 9,000 feet

    
The agent's request. Please quote one Sorensen SGI supply, 100 V and 15 kW each, with input option D (342 to 440 VAC) and no remote option. The test needs 100 V and 130 A at a site of 9,000 feet. Together they give 15 kW and 150 A. It is ready to send.

    
Verdict: Blocked.

    
    
      
  • Needed power above what the supply gives at this altitude. The test needs 13,000 W but the supply gives 9,000 W after a 40 percent altitude derating.
    • 
    

  

  

    
Narration transcript
Click to expand
    
The text of the narrated walkthrough, step by step.

    
    
      
  1. Welcome to a demo of Predictika's Validator. An AI sales agent prepares an order for Sorensen SG Series programmable DC power supplies for a test bench. Before it goes out, the agent sends it to Predictika, which checks it against a declarative model of the SG Series datasheet and returns a verdict, every reason and the figures it computed.
    2. 
      
  2. I will start with the first request by the agent.
    3. 
      
  3. The agent says. Please quote two Sorensen SGI supplies, 330 V and 30 kW each, with input option E (396 to 528 VAC) and Ethernet and RS-232C. The test needs 300 V and 170 A at a site of 3,000 feet. The rest of the request is on screen.
    4. 
      
  4. Before it goes out, the agent calls the validator.
    5. 
      
  5. Blocked, with one finding and one approval, all in one answer. The CV modification is 400 Hz AC input at 208 VAC, which needs input option C.
    6. 
      
  6. The validator returned one blocking finding and one approval in one answer. A check that reads one field at a time would pass most of this; it shows only when the choices in the order are checked against each other.
    7. 
      
  7. What this shows: for options that depend on the chassis, the control type or the input voltage (dust filter, 400 Hz, shaft locks), a rules-based CPQ is error-prone for this.
    8. 
      
  8. Now the agent resubmits.
    9. 
      
  9. The corrected request. It removes the 400 Hz input modification.
    10. 
      
  10. It calls the validator again.
    11. 
      
  11. This time the validator clears it.
    12. 
      
  12. Both earlier findings are cleared, and the validator's figure matches the order.
    13. 
      
  13. Two problems caught in one call, fixed in one pass, and a clean order. That is the value of Predictika's validator: catching mistakes before they reach the customer. Now you can try the other scenarios yourself.
    14.